EUROPEAN COMMISSION DG Competition

(Only the English text is authentic)

MERGER PROCEDURE

REGULATION (EC) 139/2004

Article 8(2) Regulation (EC) 139/2004

Date: 18/01/2018

This text is made available for information purposes only. A summary of this decision is published in all EU languages in the Official Journal of the European Union.

Parts of this text have been edited to ensure that confidential information is not disclosed; those parts are enclosed in square brackets.

EUROPEAN COMMISSION

Brussels, 18.1.2018 C(2018) 167 final

Public Version

COMMISSION DECISION

of 18.1.2018

(Only the English text is authentic)

COMMISSION DECISION

of 18.1.2018

(Only the English text is authentic)

TABLE OF CONTENTS

1. THE PARTIES ............................................................................................................. 8

2. THE OPERATION AND THE CONCENTRATION ................................................. 9

3. UNION DIMENSION ................................................................................................. 9

4. PROCEDURE .............................................................................................................. 9

5. MARKET DEFINITION ........................................................................................... 11

5.1. Introduction ................................................................................................................ 11

5.2. Product market definition ........................................................................................... 11

5.2.1. Semiconductors for automotive applications ............................................................. 11

5.2.1.1. Notifying Party's views .............................................................................................. 11

5.2.1.2. The results of the market investigation and the Commission's assessment ............... 13

5.2.2. Semiconductors for Internet of Things applications .................................................. 16

5.2.2.1. Notifying Party's views .............................................................................................. 16

5.2.2.2. The results of the market investigation and the Commission's assessment ............... 17

5.2.3. Semiconductors for mobile devices ........................................................................... 18

5.2.3.1. Baseband chipsets ...................................................................................................... 18

A) Notifying Party’s views.............................................................................................. 19

B) The results of the market investigation and the Commission's assessment ............... 21

a) Non-cellular v cellular connectivity ..................................................................................... 21

b) Standalone v integrated baseband chipsets .......................................................................... 22

c) Segmentation by cellular standard ....................................................................................... 23

d) Merchant v captive sales ...................................................................................................... 26

e) Conclusion............................................................................................................................ 27

5.2.3.2. Near Field Communication/Secure Element technology ........................................... 27

A) NFC chips................................................................................................................... 29

1. Notifying Party’s views.............................................................................................. 29

2. The results of the market investigation and the Commission’s assessment............... 31

a) Substitutability between NFC and BTLE ............................................................................ 32

b) Substitutability between NFC and QR codes....................................................................... 33

c) Substitutability between NFC and MST .............................................................................. 34

d) Installed base of POS terminals ........................................................................................... 35

e) Merchant v captive sales ...................................................................................................... 36

f) Conclusion ............................................................................................................................ 36

B) Secure Element........................................................................................................... 36

1. Notifying Party’s views.............................................................................................. 36

2. The results of the market investigation and the Commission’s assessment............... 38

2

a) Merchant v captive sales ...................................................................................................... 42

b) Conclusion ........................................................................................................................... 43

C) Combined NFC/SE solutions ..................................................................................... 43

5.2.3.3. Mobile Audio – speech enhancement software and smart amplifiers........................ 45

A) Speech enhancement software ................................................................................... 45

1. Notifying Party’s views.............................................................................................. 45

2. The results of the market investigation and the Commission's assessment ............... 46

B) Smart amplifiers ......................................................................................................... 46

1. Notifying Party's views .............................................................................................. 46

2. The results of the market investigation and the Commission's assessment ............... 47

5.2.4. Transit service technologies ....................................................................................... 47

5.2.4.1. Notifying Party's views .............................................................................................. 48

5.2.4.2. The results of the market investigation and the Commission's assessment ............... 49

5.2.5. Intellectual Property (SEPs and non-SEPs) ............................................................... 51

5.2.5.1. Notifying Party's views .............................................................................................. 52

5.2.5.2. The results of the market investigation and the Commission's assessment ............... 54

5.3. Geographic market definition .................................................................................... 55

5.3.1. All relevant semiconductor products (automotive semiconductors, IoT semiconductors, baseband chipsets, NFC, SE, and combined NFC/SE chips, smart amplifiers) .................................................................................................................. 55

5.3.1.1. Notifying Party's views .............................................................................................. 55

5.3.1.2. The results of the market investigation and the Commission’s assessment............... 56

5.3.2. Speech enhancement software ................................................................................... 56

5.3.2.1. Notifying Party's views .............................................................................................. 56

5.3.2.2. The results of the market investigation and the Commission's assessment ............... 56

5.3.3. Transit service technologies ....................................................................................... 57

5.3.3.1. Notifying Party's views .............................................................................................. 57

5.3.3.2. The results of the market investigation and the Commission's assessment ............... 57

5.3.4. Intellectual Property (SEPs and non-SEPs) ............................................................... 57

5.3.4.1. Notifying Party's views .............................................................................................. 57

5.3.4.2. The results of the market investigation and the Commission's assessment ............... 58

6. COMPETITIVE ASSESSEMENT – HORIZONTAL NON-COORDINATED EFFECTS ................................................................................................................... 58

6.1. Analytical framework................................................................................................. 58

6.2. Semiconductors for automotive applications ............................................................. 60

6.2.1. Introduction ................................................................................................................ 60

6.2.2. Infotainment MPUs .................................................................................................... 60

6.2.2.1. Notifying Party's views .............................................................................................. 60

6.2.2.2. The results of the market investigation and the Commission's assessment ............... 60

6.2.3. Infotainment radio / audio .......................................................................................... 62

6.2.3.1. Notifying Party's views .............................................................................................. 62

6.2.3.2. The results of the market investigation and the Commission's assessment ............... 62

6.2.4. Infotainment connectivity .......................................................................................... 63

6.2.4.1. Notifying Party's views .............................................................................................. 63

6.2.4.2. The results of the market investigation and the Commission's assessment ............... 63

6.2.5. Automotive chips based on non-cellular V2X technology ........................................ 64

6.2.5.1. Notifying Party's views .............................................................................................. 65

6.2.5.2. The results of the market investigation and the Commission's assessment ............... 65

A) Non-cellular V2X technologies would be important for autonomous driving and advanced safety systems ............................................................................................ 66

B) Qualcomm and NXP are not close competitors and alternative operators will remain active .......................................................................................................................... 67

C) Market entry ............................................................................................................... 68

D) Qualcomm will continue to invest on both cellular and non-cellular V2X technologies, after the Transaction ............................................................................ 69

6.2.5.3. Conclusion.................................................................................................................. 70

6.3. Semiconductors for IoT applications ......................................................................... 70

6.3.1. Introduction ................................................................................................................ 70

6.3.2. Bluetooth .................................................................................................................... 71

6.3.2.1. Notifying Party's views .............................................................................................. 71

6.3.2.2. The results of the market investigation and the Commission's assessment ............... 71

6.4. Mobile Audio ............................................................................................................. 73

6.4.1. Introduction ................................................................................................................ 73

6.4.2. Speech enhancement software ................................................................................... 73

6.4.2.1. Notifying Party's views .............................................................................................. 73

6.4.2.2. The results of the market investigation and the Commission's assessment ............... 73

6.4.3. Smart amplifiers ......................................................................................................... 76

6.4.3.1. Notifying Party's views .............................................................................................. 76

6.4.3.2. The results of the market investigation and the Commission's assessment ............... 76

7. COMPETITIVE ASSESSMENT – CONGLOMERATE NON-COORDINATED EFFECTS ................................................................................................................... 78

7.1. Analytical framework................................................................................................. 80

7.2. Market shares ............................................................................................................. 81

7.2.1. Market shares for baseband chipsets .......................................................................... 81

7.2.2. Market shares with regard to NFC chips and SE chips .............................................. 82

7.3. Market power ............................................................................................................. 84

7.3.1. Market power of the merged entity with regard to baseband chipsets ....................... 84

7.3.1.1. Notifying Party's views .............................................................................................. 84

7.3.1.2. The results of the market investigation and the Commission's assessment ............... 85

R&D activities .......................................................................................................................... 87

Certification and relationships with OEMs and MNOs ........................................................... 87

Importance for suppliers to supply chipsets supporting a variety of standards ....................... 88

Conclusion ................................................................................................................................ 89

7.3.2. Market power of the merged entity with regard to NFC, SE chips and combined NFC/SE solutions ....................................................................................................... 89

7.3.2.1. Notifying Party's views .............................................................................................. 89

7.3.2.2. The results of the market investigation and the Commission's assessment ............... 89

Market shares ........................................................................................................................... 90

Competitive pressure from suppliers of NFC chips, SE chips and "mix-and-match" solutions .................................................................................................................................... 90

7.3.3. Market power of the merged entity in relation to transit services technologies ........ 92

7.3.3.1. Notifying Party's views .............................................................................................. 92

7.3.3.2. The results of the market investigation and the Commission's assessment ............... 93

7.3.4. Market power with regard to IP (SEPs and non-SEPs)............................................ 100

7.3.4.1. Notifying Party's views ............................................................................................ 100

7.3.4.2. The results of the market investigation and the Commission's assessment ............. 101

7.4. Conglomerate effects in relation to Qualcomm's baseband chipsets and NXP's NFC and SE chips ............................................................................................................. 102

7.4.1. Introduction .............................................................................................................. 102

7.4.2. Mixed bundling and raising MIFARE royalties....................................................... 103

7.4.2.1. Background .............................................................................................................. 103

7.4.2.2. Analysis of mixed bundling ..................................................................................... 104

A) Notifying Party’s views............................................................................................ 106

B) Commission's assessment ........................................................................................ 108

1. Ability to engage in mixed bundling ........................................................................ 108

2. Incentive to engage in mixed bundling .................................................................... 115

3. Likely effects on competition................................................................................... 125

Effects of mixed bundling ...................................................................................................... 126

Effects of increased royalties for MIFARE in addition to mixed bundling ........................... 129

7.4.3. Pure bundling and tying ........................................................................................... 133

7.4.3.1. Background .............................................................................................................. 133

7.4.3.2. Analysis of pure bundling and tying ........................................................................ 134

A) The Notifying Party's view ...................................................................................... 134

1. Ability ...................................................................................................................... 134

2. Incentive ................................................................................................................... 136

3. Effects ...................................................................................................................... 137

B) Commission's assessment ........................................................................................ 138

1. Ability to engage in pure bundling or tying ............................................................. 138

2. Incentive to engage in pure bundling or tying ......................................................... 140

3. Likely effects on competition................................................................................... 144

7.4.4. Degradation of interoperability ................................................................................ 145

7.4.4.1. The Notifying Party's views ..................................................................................... 146

7.4.4.2. Commission's assessment ........................................................................................ 147

A) Ability to engage in degradation of interoperability ................................................ 147

B) Incentive to engage in degradation of interoperability ............................................ 149

C) Likely effects on competition................................................................................... 151

7.4.5. Conclusion on conglomerate effects in relation to Qualcomm's baseband chipsets and NXP's NFC and SE chips ......................................................................................... 153

7.5. Conglomerate effects related to IP licensing of NFC technology............................ 153

7.5.1. Preliminary remarks ................................................................................................. 154

7.5.2. The Parties' pre-merger IP licensing practices ......................................................... 156

7.5.2.1. NFC's licensing practices ......................................................................................... 157

7.5.2.2. Qualcomm's licensing practices ............................................................................... 157

A) Qualcomm's wireless patent portfolio ...................................................................... 157

B) Qualcomm's licensing model ................................................................................... 158

1. Qualcomm's device-level licensing and "no license-no chip" practices .................. 158

2. Qualcomm's portfolio-based license and royalty ..................................................... 158

3. Qualcomm's licensees .............................................................................................. 159

4. Renegotiations of licensing terms ............................................................................ 160

7.5.3. The Notifying Party's view ...................................................................................... 160

7.5.4. Results of the phase I and phase II market investigations ....................................... 161

7.5.5. The Commission's assessment ................................................................................. 161

7.5.5.1. Impact of the Transaction on the merged entity's bargaining power and resulting royalties .................................................................................................................... 161

A) The merged entity's IP position ................................................................................ 161

B) Impact on the merged entity's bargaining position and royalties or licensing fees .. 163

C) Impact on royalties in the context of Qualcomm's broad portfolio licenses ............ 165

7.5.5.2. The counterarguments put forward by the Notifying Party do not alter the Commission’s findings............................................................................................. 166

A) The merged entity would be able to charge higher royalties in the long run despite Qualcomm’s pre-existing portfolio licenses ............................................................ 166

B) The increase in royalties for NFC patents is likely to be significant ....................... 167

C) The alleged historic nominally stable royalty rate does not alter the Commission’s findings..................................................................................................................... 168

7.5.6. Conclusion on IP licensing of NFC technology ....................................................... 171

7.6. Conglomerate effects in relation to Qualcomm's baseband chipsets, NXP's NFC and SE chips as well as licensing of the merged entity's IP ........................................... 171

7.6.1. Introduction .............................................................................................................. 171

7.6.2. Notifying Party’s views............................................................................................ 172

7.6.3. Third parties' views .................................................................................................. 174

7.6.4. Commission's assessment ........................................................................................ 174

7.6.4.1. Preliminary remarks ................................................................................................. 174

7.6.4.2. Ability to extend the NLNC strategy to NXP’s NFC/SE chips ............................... 175

7.6.4.3. Incentive to extend the NLNC strategy to NXP’s NFC/SE chips ............................ 175

7.6.4.4. Likely effects of extending the NLNC strategy to NXP’s NFC/SE chips ............... 177

7.6.5. Conclusion on conglomerate effects in relation to Qualcomm's baseband chipsets, NXP's NFC and SE chips as well as licensing of the merged entity's IP ................. 177

8. COMMITMENTS .................................................................................................... 178

8.1. Analytical framework............................................................................................... 178

8.2. Procedure.................................................................................................................. 179

8.3. Commission's assessment of the First Commitments .............................................. 179

8.3.1. Description of the First Commitments ..................................................................... 179

8.3.2. Results of the market test of the First Commitments ............................................... 182

8.3.3. Assessment of the First Commitments ..................................................................... 186

8.3.3.1. Carve-out and non-assert remedies .......................................................................... 186

8.3.3.2. Interoperability remedy ............................................................................................ 188

8.3.3.3. MIFARE remedy ...................................................................................................... 189

8.3.3.4. Conclusion................................................................................................................ 192

8.4. Commission's assessment of the Final Commitments ............................................. 192

8.4.1. Description of the Final Commitments .................................................................... 192

8.4.2. Assessment of the Final Commitments .................................................................... 195

8.4.2.1. Carve-out and non-assert remedies .......................................................................... 195

8.4.2.2. Interoperability remedy ............................................................................................ 197

8.4.2.3. MIFARE remedy ...................................................................................................... 200

8.4.2.4. Conclusion................................................................................................................ 201

8.5. Conclusion................................................................................................................ 202

9. CONDITIONS AND OBLIGATIONS .................................................................... 202

THE EUROPEAN COMMISSION,

Having regard to the Treaty on the Functioning of the European Union,

Having regard to the Agreement on the European Economic Area, and in particular Article 57 thereof,

Having regard to Council Regulation (EC) No 139/2004 of 20 January 2004 on the control of concentrations between undertakings, and in particular Article 8(2) thereof,

Having regard to the Commission's Decision of 9 June 2017 to initiate proceedings in this case,

Having regard to the opinion of the Advisory Committee on Concentrations,

Having regard to the final report of the Hearing Officer in this case ,

Whereas:

(1) On 28 April 2017 the Commission received a notification of a proposed concentration pursuant to Article 4 of Council Regulation (EC) No 139/2004 ("the Merger Regulation") by which the undertaking Qualcomm Incorporated (United States of America), through its indirect wholly owned subsidiary Qualcomm River Holdings B.V. (the Netherlands) (together referred to as "Qualcomm" or the "Notifying Party") would acquire within the meaning of Article 3(1)(b) of the Merger Regulation control of NXP Semiconductors N.V. ("NXP", the Netherlands) by way of a purchase of shares (the "Transaction"). Qualcomm and NXP are jointly referred to as the "Parties".

1 OJ L 24, 29.1.2004, p. 1 ("the Merger Regulation"). With effect from 1 December 2009, the Treaty on the Functioning of the European Union ("TFEU") has introduced certain changes, such as the replacement of "Community" by "Union" and "common market" by "internal market". The terminology of the TFEU will be used throughout this decision.

2 OJ C ...,...200. , p....

3 OJ C ...,...200. , p....

4 Publication in the Official Journal of the European Union No C 143, 06.05.2017, p. 6.

Qualcomm’s IP portfolio includes thousands of standard essential patents ("SEPs")

(5) NXP is active in the manufacturing and sale of semiconductors, in particular ICs and single unit (discrete) semiconductors. NXP sells High Performance Mixed Signal ("HPMS") devices, which comprise application-specific semiconductors and system solutions for the following segments, namely (i) Automotive; (ii) Secure Identification Solutions; (iii) Secure Connected Devices; and (iv) Secure Interfaces and Power.

2. THE OPERATION AND THE CONCENTRATION

(6) On 27 October 2016, Qualcomm Incorporated, through its indirect, wholly-owned Dutch subsidiary Qualcomm River Holdings B.V., entered into a purchase agreement with NXP, pursuant to which Qualcomm will commence a tender offer to acquire all of the issued and outstanding common shares of NXP, thus acquiring sole control of NXP.

(7) The Transaction therefore constitutes a concentration pursuant to Article 3(1)(b) of the Merger Regulation.

(8) The Notifying Party explains that the Transaction is a key step in its strategy to diversify away from the mobile device sector, where it holds particular expertise in cellular and connectivity technologies, to areas in which it currently has no or limited presence. The Notifying Party explains that the Transaction would accelerate the development of smart solutions in particular in the automotive, industrial, smart home and healthcare areas.

3. UNION DIMENSION

(9) The undertakings concerned have a combined aggregate world-wide turnover of more than EUR 5 000 million (Qualcomm: EUR 21 200 million; NXP: EUR 8 650 million). Each of them has a Union-wide turnover of more than EUR 250 million (Qualcomm: EUR 712 million; NXP: EUR 1 520 million), but they do not achieve more than two-thirds of their aggregate Union-wide turnover within one and the same Member State. The notified operation therefore has a Union dimension.

4. PROCEDURE

(10) Based on the results of the phase I market investigation, the Commission raised serious doubts as to the compatibility of the Transaction with the internal market and adopted a decision to initiate proceedings pursuant to Article 6(1)(c) of the Merger Regulation on 9 June 2017.

(11) The Notifying Party submitted its written comments to the Article 6(1)(c) decision on 28 June 2017.

5 Patents that are essential to a standard are those that cover technology to which a standard makes reference and that implementers of the standard must use in standard-compliant products. These patents are known as SEPs. SEPs differ from non-essential patents ("non-SEPs"), which can generally be designed around to comply with a standard.

6 Turnover calculated in accordance with Article 5 of the Merger Regulation and the Commission Consolidated Jurisdictional Notice (OJ C 95, 16.4.2008, p. 1).

(12) During the phase II investigation, the Commission sent several requests for information (“RFIs”) to the Parties and to third party market participants. In particular, the Commission sent market investigation questionnaires to customers and competitors of semiconductors for the automotive industry, customers and competitors in the Internet of Things ("IoT") sector, device original equipment manufacturers ("OEMs"), to suppliers of baseband chipsets, of Near Field Communication ("NFC") and Secure Element ("SE") technology chips, to mobile network operators ("MNOs"), and to public transit authorities and reader infrastructure integrators. The Commission also sent targeted requests for information to device OEMs, to suppliers of baseband chipsets, and to suppliers of NFC and SE chips.

(13) On 28 June 2017, the Commission adopted a decision pursuant to Article 11(3) of the Merger Regulation, compelling the Notifying Party to supply the information which it had requested on 14 June 2017 with RFI 18 pursuant to Article 11(2) of the Merger Regulation, which the Notifying Party had failed to submit within the time limit fixed by the Commission. The decision also suspended the time limit laid down in Article 10(3) of the Merger Regulation until the end of the day the Commission would receive the required information.

(14) On 16 August 2017, the Notifying Party submitted a response to RFI 18 and the suspension of the time limit expired at the end of that day.

(15) On 5 September 2017, the Commission adopted a decision pursuant to Article 11(3) of the Merger Regulation, compelling the Notifying Party to supply the information which it had requested on 14 June 2017 with RFI 20 pursuant to Article 11(2) of the Merger Regulation, which the Notifying Party had failed to submit within the time limit fixed by the Commission. The decision also suspended the time limit referred to in Article 10(3) of the Merger Regulation as of 17 August 2017 until the end of the day the Commission would receive the required information.

(16) On 4 October 2017, the Commission adopted a decision pursuant to Article 11(3) and Article 15 of the Merger Regulation, compelling the Notifying Party to supply certain information and documents responsive to RFI 20, which had not yet been provided to the Commission and imposing a periodic penalty payment should the Notifying Party fail to supply the information requested within the period prescribed.

(17) On 5 October 2017, the Notifying Party proposed formal commitments to eliminate the Commission’s findings that the Transaction would give rise to a significant impediment to effective competition. The Commission launched the market test for the commitments on 6 October 2017.

(18) On 17 November 2017, the Notifying Party completed its response to RFI 20 and the suspension of the time limit expired at the end of that day.

(19) Taking into account the Commission’s comments and the feedback from the market test, the Notifying Party subsequently submitted a final set of commitments on 10 November 2017.

(20) The Advisory Committee discussed the draft of this Decision on 8 January 2018 and issued a favourable opinion.

5. MARKET DEFINITION

5.1. Introduction

(21) The Transaction concerns the semiconductor industry.

(22) Semiconductors are materials, such as silicon, which can act as an insulator, but are also capable of conducting electricity. Semiconductors are at the heart of devices such as diodes, transistors and other electronic components, and can be found in virtually every electronic device today.

(23) Semiconductor devices are rarely bought as end-products by consumers. They are mainly bought by equipment manufacturers in virtually all sectors within the electronic equipment industry.

(24) The Transaction concerns different industry sectors, that can be broadly categorised in three macro areas, namely semiconductors for automotive applications, semiconductors for IoT applications, and semiconductors for mobile devices including in particular baseband chipsets, NFC and SE technology, and mobile audio products (speech enhancement software and smart amplifier chips). The Transaction also concerns transit service technology used for contactless public transport ticketing and fare collection by means of mobile devices and the IP related to baseband chipset and NFC technology.

5.2. Product market definition

5.2.1. Semiconductors for automotive applications

(25) In the automotive segment, Qualcomm is mainly active in repurposed chips primarily developed for the mobile segment, including the Snapdragon line of baseband processors. Qualcomm's automotive sales accounts for [0-5]% of Qualcomm's annual turnover. Qualcomm's product offerings in the automotive segment include the following product groups: infotainment, connectivity, and Advanced Driver Assistance Systems ("ADAS").

(26) NXP is a leading automotive semiconductor supplier; it has a sales share of approximately [10-20]% of the overall automotive semiconductor industry and its automotive sales account for a third of NXP's annual turnover. NXP provides solutions for applications such as infotainment, ADAS, chassis & safety and body and comfort (for example lighting).

5.2.1.1. Notifying Party's views

(27) The Notifying Party submits that there are very few overlaps in the Parties' activities in the automotive space and that the industry for semiconductors in the automotive space can be segmented by (i) semiconductor type; or (ii) by field of application/end-use.

(28) With regard to the segmentation by semiconductor type, in line with previous Commission decisions, four main categories of semiconductor devices can be identified: (i) ICs (also known as microchips or chips); (ii) discretes; (iii) optical semiconductors; and (iv) sensors and actuators. Furthermore, ICs can be distinguished on the basis of whether they incorporate digital or analog technology,

8 At the Advisory Committee all present Member States agreed that that the Transaction must be declared compatible with the internal market in accordance with Articles 2(2) and 8(2) of the Merger Regulation.

9 Commission decision of 17 September 2015 in Case M.7585 – NXP Semiconductors/ Freescale Semiconductor.

Qualcomm’s IP portfolio includes thousands of standard essential patents ("SEPs")

Automotive infotainment is a collection of hardware and software in automobiles that provides audio and/or video entertainment.

Semiconductors for automotive infotainment systems could be further segmented in (i) radio/audio; (ii) connectivity (for example Wi-Fi, Bluetooth); (iii) cellular/modem; (iv) graphics, display and multimedia processing (“Infotainment MPU”).

ADAS encompasses a broad range of features that enable a vehicle to “see,” “sense” and “react” to the objects that surround it, through the use (and combination) of semiconductor-driven technologies. Over time, ADAS systems are expected to evolve into more sophisticated systems and eventually autonomous driving system. Therefore, for the purposes of this decision, the Commission refers to ADAS systems as the set of technologies that can perform from driving assistance (level 1 according to the Society of Automotive Engineers ("SAE")) to full autonomous driving (level 5 according to SAE).

IEEE stands for Institute of Electrical and Electronics Engineers

In the United States, the National Highway Traffic Safety Administration ("NHTSA") proposed the adoption of a safety standard requiring all new light vehicles sold on the U.S. market to be capable of vehicle-to-vehicle communications based on the DSRC communications standard. See NHTSA’s Notice of Proposed Rulemaking on V2V Communications, 13 December 2016: https://www.nhtsa.gov/press-releases/us-dot-advances-deployment-connected-vehicle-technology-prevent-hundreds-thousands, [Doc ID 3293]

Article 6(1)(c) decision, the Notifying Party argues that non-cellular V2X systems are composed of different components, namely, non-cellular V2X radio, software, secure storage, processor and interface with the rest of the car (for example, BUS) and that the Parties overlap only in the provision of non-cellular V2X radio chip.

5.2.1.2. The results of the market investigation and the Commission's assessment

(33) The majority of respondents to the market investigation indicated that semiconductor devices can be categorised into four main distinct categories that are (i) integrated circuits ("ICs"); (ii) discretes; (iii) optical semiconductors; and (iv) sensors and actuators. All respondents further noted that ICs can be further distinguished between digital ICs and analog ICs, and that digital ICs belong to three categories, (i) microcomponents ICs; (ii) memory ICs; and (iii) logic ICs. The majority of the respondents further consider that there are three categories of microcomponents ICs, namely (i) microprocessors (“MPUs”); (ii) microcontrollers (“MCUs”); and digital signal processors (“DSPs”).

(34) With regard to a potential distinction between general purpose microcomponents ICs from application specific microcomponents ICs, the majority of respondents further believe that a distinction should be made since specific application needs specific microcomponents. Only two competitors note that it would be increasingly difficult to distinguish between the two. The majority of respondents further believe that there is no reason to depart from the Commission's previous assessment that application specific microcomponents ICs can be distinguished depending on their category of application and that application specific microcomponents ICs of one category, such as automotive, are likely not substitutable with those of another category (for example, automotive / wireless communications).

(35) With regard to semiconductors used for automotive applications, the vast majority of customers responding to the market investigation consider that different products markets should be considered for specific end-use/application for which semiconductor products are used and that there should be a distinction between several function blocks that are (i) powertrain; (ii) chassis; (iii) safety; (iv) body and comfort; (v) infotainment; and (vi) security. From the competitors' side, while most of the respondents clearly agree with this segmentation, a few respondents note that a distinction should be made for specific function blocks. One respondent notes that the proposed classification is true only when considering ASSPs/ASICs but not general purpose ICs.

(36) With regard to automotive infotainment chips, the majority of the respondents to the market investigation consider that the market segment could be segmented in the following functions blocks: (i) radio / audio; (ii) connectivity; (iii) cellular / modem; and (iv) graphics, display and multimedia processing ("Infotainment MPUs"). One respondent, in particular, considers that those are very different devices for functionality, performance and price and substitution is impractical. The majority of respondents further notes that those products should be considered as complementary to each other rather than alternatives to each other and that, to start providing semiconductors for a specific function block (for example infotainment MPUs) may require significant time and investments. Only one respondent considers the product of the different function blocks as substitutes; however, the same respondent highlights that it is not active in the market.

(37) With regard to infotainment MPUs, the majority of respondents consider that Infotainment MPU could be further distinguished into two tiers: (i) Entry/Display Audio (“DA”) tier; (ii) mid/premium-high tier. Intel and Continental note that the two systems have different displays and audio requirements. Only one respondent pointed out that it may require a higher investment to create two versions. The majority of respondents also note that to start providing semiconductors for another Infotainment MPU category would require significant time and investment. However, some customers explained that as both are built with the same architecture, less investment would be required. One competitor also said that because those products usually contain licensed processor cores from third party, license of a new core with greater capacity would be cost effective and quick. Moreover, the majority of respondents consider the automotive infotainment MPU used for the different tiers (Entry/Display Audio (“DA”) tier and mid/premium-high tier) as being complementary with each other rather than alternatives to each other.

(38) With regard to infotainment connectivity, the vast majority of respondents consider (i) Wi-Fi; (ii) Bluetooth (“BT”); (iii) Wi-Fi/BT Combo; (iv) cellular (including 3G, LTE and 5G); (v) 802.11p as being complementary with each other rather than alternatives to each other.

(39) For the purposes of this Decision, the exact product market definition with regard to chips for the infotainment automotive application can be left open as the Transaction does not raise competition concerns as regards its compatibility with the internal market irrespective of precise product market definition.

(40) With regard to chips used in the safety function block, the majority of the respondents to the market investigation consider that they could be further distinguished between (i) passive safety; (ii) active safety – ADAS. In particular, one respondent notes that passive safety is a mature field with established supply chain, whereas ADAS is a new field with a much higher level of complexity to master. However two automotive customers disagree with this distinction, and explained that for the IC, the distinction does not matter.

(41) The majority of respondents further consider that the automotive ADAS chips in the following function blocks: (i) Ultrasonic; (ii) Lidar; (iii) Radar; (iv) camera; (v) V2X communication are complementary rather than alternatives to each other since they are different technologies used for different purposes and functions. However, one respondent notes that depending on the requirement and performance of the ADAS chips, or depending on the semiconductor used in the function block, they can be substitutable or complement each other.

(42) The majority of customers also consider the automotive V2X chips based on cellular and non-cellular (for example, 802.11.p) technology as complements rather than alternatives to each other. Around half of the competitors also agree with this. In particular, some customers highlight how non-cellular is the only technology that could be used for vehicle-to-vehicle safety function in the next years, while cellular technology may be used when 5G will be developed for vehicle-to-individual safety functions. Amongst the competitors, while some consider that there are overlaps between the two technologies, around half believe that they can coexist in a vehicle. However, both customers and competitors note that the two technologies have different deployment times since the non-cellular one is already available in the market and tested by several automotive manufacturers while the cellular one is at earlier stage of development and two-three years behind.

(43) Moreover, the majority of respondents note that in order to start providing chips based on a different communication standard (for example, from cellular V2X to non-cellular V2X), a company would incur significant time and investment. One respondent notes that this would not be possible since cellular communication uses completely different technologies that 802.11.p. In particular, it would be difficult to start providing 802.11.p technology due to its high development time. Only one competitor said that a company could switch easily, but gave no further explanation. However, some respondents consider the investment for Wi-Fi chips suppliers to start developing 802.11.p technology chip to be relatively small. Others also note that Marvell, a Wi-Fi chips supplier, just launched its non-cellular radio offering.

(44) With regard to the different components of non-cellular V2X systems, the majority of respondents consider those components, namely non-cellular V2X radio, software, secure storage, processor and interface with the rest of the car (for example, BUS) to be complementary. Respondents note how all those technologies are needed to build an autonomous vehicle.

For the purposes of this Decision, the exact product market definition with regard to chips for the safety function for automotive application can be left open as the Transaction does not raise competition concerns as regards its compatibility with the internal market irrespective of precise product market definition.

5.2.2. Semiconductors for Internet of Things applications

The IoT refers to interconnected objects – wearables, appliances, industrial automation (or vehicles and infrastructure) – that have processing capabilities and that can transmit or receive data or both through different communications standards, such as Bluetooth, NFC, Wi-Fi, and cellular.

5.2.2.1. Notifying Party's views

The Notifying Party submits that in the IoT space, the Parties have a highly complementary presence and the overlaps between the Parties' activities are very limited.

Similar to semiconductors for automotive applications, the market for semiconductors for IoT applications can be segmented by (i) semiconductor type; or (ii) by field of application/end-use.

As regards the possible segmentation by semiconductor type, the Notifying Party identifies, in line with previous Commission decisions, four main categories of semiconductor devices: (i) integrated circuits (“ICs”), (ii) discretes, (iii) optical semiconductors, and (iv) sensors and actuators. Furthermore, ICs can be distinguished on the basis of whether they incorporate digital or analog technology, or a combination of both. Digital ICs can be further segmented into three categories, microcomponents, memory ICs, and logic ICs. In turn, microcomponents can be further subdivided in three types, namely microprocessors (“MPUs”), microcontrollers (“MCUs”) and Digital Signal Processors (“DSPs”), each of which can be subdivided between general purpose and application specific. Those semiconductors provide processing and connectivity functions. In terms of connectivity, many IoT devices today use the lower band short range wireless standards (usually sub-GHz to 2.4 GHz frequencies) such as Bluetooth Classic (“BT”), Bluetooth low energy (“BTLE” or “BLE”), Radio Frequency Identification (“RFID”), NFC, and IEEE 802.15.4 (which includes ZigBee and Thread), and for some other IoT applications, Wi-Fi (802.11n/802.11ac) and cellular (for example,

See responses to questions 12 of questionnaires Q15 – automotive competitors and Q16 – automotive customers.

See responses to question 7 of questionnaires Q15 – automotive competitors and Q16 – automotive customers.

Commission decision of 17 September 2015 in Case M.7585 – NXP Semiconductors/ Freescale Semiconductor.

Analog ICs can be divided between general purpose analog ICs and application specific analog ICs.

Code Division Multiple Access (“CDMA”) or the Orthogonal Frequency Division Multiple Access (“OFDMA”) family of technologies connectivity may also be used.

In the IoT space, Qualcomm and NXP supply general purpose and application specific MPUs, application specific MCUs, and application specific DSPs.

However, the Notifying Party submits that the overlaps between the Parties are limited. With regard to MPUs, the Notifying Party argues that only NXP is present in the segment for general purpose MPUs considering that Qualcomm has only recently released a general purpose MPU. In the segment for application specific MPUs, Qualcomm is active through its application processors ("AP") (Snapdragon product line) designed for mobile devices as well as repurposed Snapdragon APs intended for use in for example wearables. NXP supplies its 'i-MX' line of MPUs in this segment. With regard to MCUs, there is an overlap for application specific MCUs, in particular connectivity chips, and more specifically connectivity chips implementing the Bluetooth low energy (BTLE) standard. As for DSPs, the Notifying Party submits that the overlap is due to the DSP functionality of Qualcomm's baseband chips, while NXP supplies stand-alone DSPs, including Near Field Magnetic Induction ("NFMI") based chips used in IoT applications.

As regards the field of application, the Notifying Party submits that semiconductors used in the IoT space can also be segmented by field of application/end-use, with one such categorisation being hearables (including true wireless stereo ("TWS") headphones/speakers), wearables (including smart glasses, smart watches, smart trackers, body sensors, wearable cameras), smart homes (including smart routers, home automation, entry management, white appliances, LED lighting, media devices), drones and smart cameras, and smart cities (energy & metering, infrastructure, industrial/building, transportation).

The Notifying Party submits that both Parties provide solutions for TWS headphones, speakers, wearables, drones, smart lighting and smart appliances, but that their products serve different purposes within those categories.

5.2.2.2. The results of the market investigation and the Commission's assessment

All respondents to the market investigation confirmed that semiconductors may be divided into the four main categories (ICs, discretes, optical semiconductors and sensors and actuators). Respondents also agreed that within ICs a distinction should be made between MPUs, MCUs and DSPs. The vast majority of respondents consider that connectivity chips have a distinct functionality from other semiconductors used in IoT (MCU, MPU). Furthermore, respondents note that connectivity chips may be further segmented in BT, BTLE, Wi-Fi, NFC and IEEE 802.15.4.

Finally, overall, respondents to the market investigation also subscribe to the segmentation of the market by end-use proposed by the Notifying Party, although additional categories are also suggested. With regard to the segmentation by field use-application, the market investigation provided mixed results. While most of the customers consider that a distinction should, at least, be made for semiconductors used in the following function blocks: (i) hearables (including true wireless headsets), (ii) wearables (watches and fitness), (iii) drones and smart cameras, (iv) smart cities, and (v) smart home (including home entertainment), competitors' replies were mixed. In particular, Renesas and Nordic highlight that the market for standardised chips is generally a single market, and while customers can choose to put the chips in different end applications, the chips and the market for their sales is the same. With regard to a potential segmentation of semiconductors in the hearables IoT space, the market investigation did not provide a clear indication on whether TWS headphones provide different functionalities respect to other semiconductors in the hearables IoT space.

For the purposes of this Decision, the exact product market definition with regard to IoT semiconductors can be left open as the Transaction does not raise competition concerns as regards its compatibility with the internal market, irrespective of the precise product market definition.

5.2.3. Semiconductors for mobile devices

5.2.3.1. Baseband chipsets

For the purpose of providing mobile cellular connectivity, mobile devices rely on a baseband processor, which enables the connection of mobile devices to mobile telecommunication networks.

A baseband processor/modem is typically paired with two additional components to complete its functionality: the Radio Frequency ("RF") integrated circuit (or "RF transceiver") and the Power Management ("PM") integrated circuit ("PMIC"). All three functionalities (baseband processor/modem, RF transceiver and PMIC) are necessary for mobile cellular connectivity. Combined, the three components are referred to as a "baseband chipset" ("BC"). Although each component can be sourced individually, they are usually sourced together.

Baseband chipsets are sold either on a standalone basis ("standalone baseband chipset" or "standalone BC") or combined with an application processor ("AP") ("integrated baseband chipsets" or "integrated BC"). The AP runs the operating system and applications of mobile devices.

Baseband chipsets, whether in standalone or integrated mode, implement one or multiple cellular standards from the same or from different technology families and generations. A baseband chipset might implement only one standard (so called single-mode baseband chipset), or a combination of several standards, with later generation baseband chipsets often being backward compatible with earlier standards (so-called multi-mode baseband chipsets).

Qualcomm supplies baseband chipsets. NXP does not. Qualcomm supplies both integrated and standalone baseband chipsets, and supplies the three components (baseband chip/modem, RF transceiver, and PMIC) both independently and combined.

A) Notifying Party’s views

The Notifying Party considers that the relevant baseband chipset market should include chipsets supporting all wireless communication standards, including cellular (for example, GMS, CDMA, UMTS, LTE) and non-cellular (for example, Wi-Fi, WiMAX) standards. The Notifying Party argues that cellular and non-cellular chipsets compete against each other, exert competitive constraints on each other and are substitutable both from a demand and supply side perspective.

The Notifying Party considers that different iterations of the same standard constrain the pricing of other iterations of that same standard, and on the other hand, chipsets supporting other wireless standards constrain the pricing of chipsets supporting a particular cellular standard, and vice versa. The Notifying Party further argues that from the demand side, customers/end-users consider Wi-Fi and cellular wireless communication networks interchangeably, as Wi-Fi coverage is increasingly available in both public and private places, including through "hotspot" networks. On the supply side, and with particular reference to the standardised nature of the technologies incorporated in baseband chipsets, and specifications being publicly available, market entry is facilitated and companies active in a specific type of wireless communication standard are able to develop in the short/medium term solutions that extend or add other wireless communication standards to their product offering.

The three components are usually implemented on separate pieces of silicon and packaged into separate chips, but can also be packaged in the same chip as the baseband processor.

The Global System for Mobile Communications, or GSM, is a cellular communication standard developed by the European Telecommunication Standards Institute (ETSI) to describe technologies for second generation (2G) digital cellular networks. The Universal Mobile Telecommunications System, or UMTS, is a third generation (3G) wireless and mobile communications standard developed by ETSI and the 3rd Generation Partnership Project (3GPP) with capabilities going beyond 2G standards in supporting multimedia services. The UMTS standard has over time evolved to provide improved characteristics such as higher data rates of broadband connectivity. Long-Term Evolution, or LTE, was developed by 3GPP, further increasing the capacity and speed of data. It is commonly referred to as fourth generation (4G) standard.

See the Notifying Party's response of 31 March 2017 to the Commission's request for information ("RFI") of 28 February 2017.

The Notifying Party argues against a further segmentation of baseband chipsets based on the type of devices (smartphones versus other electronic devices such as tablets), by the type of smartphone (high-end versus low-end smartphones), by cellular technology (GSM, UMTS, CDMA, LTE), and by functionality (distinguishing between stand-alone and integrated baseband chipsets). The Notifying Party considers that in any event, the exact product market definition can be left open, as the Transaction will not result in any affected market for baseband chipsets.

In response to the Article 6(1)(c) decision the Notifying Party argues that while every iteration of a given standard and, more broadly, technology generation, paves the way for the next standard to allow for a phased evolution and elimination of possible performance issues, a segmentation of the market by standard results in an arbitrary distinction within a chain of substitution and groups together technologies with very different performance characteristics. There is furthermore a supply-side substitutability between suppliers of baseband chipsets implementing the UMTS standard and chipsets implementing the LTE standard with an overwhelming majority of suppliers that have both the technological know-how and do not require a license to produce chipsets compliant with the various technologies.

The Notifying Party also considers that both "merchant" and "captive" sales should be included in the relevant product market for baseband chipsets; device OEMs' "captive" sales constrain "merchant" sales as vertically integrated customers are able to divert their demand to baseband chipsets developed in-house and to offer that capacity to third party customers.

In response to the Article 6(1)(c) decision the Notifying Party further argues that the "captive" suppliers are given recurrent opportunities to enter the "merchant" market and such vertically integrated device OEMs can "toggle" between in-house and external supply from one device to another. The Notifying Party believes that in-house suppliers, such as Samsung, are neither capacity constrained, nor do their baseband chips lag behind technology-wise to those available on the merchant market, as evidenced by the fact that both Samsung and Huawei use their own production for some of their premium mobile devices. In addition, Samsung also supplies (limited volumes of) baseband chipsets on the merchant market, which is evidence supporting the notion that "merchant" and "captive" sales indeed form part of the same relevant market. The fact that the market analysis report by Linley Group forecasts that both Samsung and Huawei will source approximately [40-50]% of their baseband chipset demand in-house shows that "captive" sales constitute a competitive constraint on "merchant" sales.

The Notifying Party considers that instead of segmenting by cellular standard nomenclature, such as UMTS or LTE, a more appropriate segmentation would be by generation of standards, namely 2G, 2.5G, 3G, 4G, and 5G. This would also be more in line with the Commission's previous decisional practice where an analysis of the market for telecommunication network equipment was made with reference to cellular technology generations (see Commission decision of 24 July 2015 in case M.7632 – Nokia / Alcatel-Lucent).

See response by the Notifying Party of 28 June 2017 to the Commission decision of 9 June 2017 to initiate proceedings pursuant to Article 6(1)(c) of the Merger Regulation ('Article 6(1)(c) decision'), [Doc ID 1331].

The Commission considers that the relevant product market consists of standalone and integrated baseband chipsets, segmented by cellular standard (LTE, UMTS, CDMA, GSM) with a distinction between single-mode and multi-mode baseband chipsets. Single-mode chipsets support only one cellular standard, while multi-mode chipsets support two or more cellular standards, usually earlier generation cellular standards still in use. With regard to chipsets implementing the LTE cellular standard, a distinction is made between single-mode LTE chipsets and multi-mode LTE chipsets, where single-mode LTE chipsets do not form part of the same product market as multi-mode LTE chipsets.

a) Non-cellular v cellular connectivity

Chipsets supporting non-cellular wireless communication standards such as Wi-Fi or WiMAX are not substitutable for cellular chipsets, in particular multi-mode LTE baseband chipsets, namely chipsets implementing the LTE cellular standard but also earlier generation standards (in particular UMTS and GSM).

Wi-Fi access is not readily available to smartphone users wherever they are. This is because Wi-Fi is a wireless local area network (WLAN) and mostly limited to homes, work places and "hotspots". Users cannot, therefore, rely on Wi-Fi while on the move and outside "hotspot" areas. Users can connect with their smartphones to Wi-Fi and receive and send data whenever in such a local area network, but cannot rely on it elsewhere or as the primary connectivity technology for their smartphones.

The argument by the Notifying Party that customers and end-users consider Wi-Fi and cellular networks to be interchangeable, giving the example of Apple marketing both a non-cellular (Wi-Fi-only) and a cellular/non-cellular (UMTS/LTE together with Wi-Fi) version of the iPad tablet computer as evidence of the demand-side substitutability of these technologies, is not convincing. Tablets with cellular connectivity represent only a small part of cellular mobile devices taken as a whole. According to The Linley Group, voice-over-Wi-Fi has been […] .

Similarly for WiMAX, its coverage is limited and cannot, therefore, be relied upon by users for the purposes of connectivity. In 2015, less than 20% of Union households were covered by WiMAX for broadband access (with huge disparities by country, with a majority of Member States either having no WiMAX coverage or less than 5% in 2015), compared to approximately 86% coverage for the cellular LTE technology.

"Multi-mode LTE baseband chipset" denotes a baseband chipset implementing LTE in combination with earlier generation cellular standards, in particular UMTS and GSM.

For example, The Linley Group indicates that smartphones count for a majority of all mobile devices shipped, with tablet shipments on the decline, especially high-end tablets such as the Apple iPad "falling rapidly". According to The Linley Group, tablets (both cellular and non-cellular versions, including e-readers) account for only some [5-10]% of the total amount of smartphones, mobile phones and tablets (including e-readers) shipped in 2017. The Linley Group estimates that by 2020, [30-40]% of tablets will have cellular baseband chips, up from [20-30]% in 2015; The Linley Group: "Mobile Semiconductor Market Share Forecast 2015-2020", pp. 5-6, 13-14; Form CO, Annex 4.16 [Doc ID: 327].

See The Linley Group: "Mobile Semiconductor Market Share Forecast 2015-2020", page 10; Form CO, Annex 4.16 [Doc ID: 327].

The market investigation indicates that non-cellular wireless connectivity standards (such as Wi-Fi, WiMAX) do not constitute viable alternatives to cellular connectivity standards (such as LTE, UMTS), and are not substitutable either from a demand or supply side perspective. Respondents consider that non-cellular and cellular connectivity standards have different characteristics, especially in coverage where cellular connectivity allows for wider coverage than non-cellular standards such as Wi-Fi; this is especially crucial for mobile phones where cellular technology remains the main connectivity technology. Cellular connectivity has, as explained by one device OEM "wide coverage and high mobility" while non-cellular connectivity, such as Wi-Fi, has "fast transmission speed and simple networking". Sequans, a company founded "to address the WiMAX market" and becoming "the acknowledged leader in WiMAX", and currently refocused on single-mode LTE baseband chipsets mainly for IoT, confirms that the two types of technologies are not viable alternatives and "[c]ellular technology can achieve/support application that non cellular could not like mobility, point to point access".

b) Standalone v integrated baseband chipsets

Smartphones require both a baseband processor (to send and receive cellular radio signals) and an application processor (to run the apps on a smartphone). Two main architectures can be deployed by a device OEM, namely one which is composed of standalone components (standalone baseband processors and standalone application processor) or alternatively an integrated architecture composed of a baseband processor and application processor on a single silicon die. The Commission considers that integrated and standalone baseband chipsets are substitutable.

Baseband processors and other components, such as application processors, are distinct products based on product characteristics, price, intended use or application. There is no demand or supply-side substitutability between baseband processors and other stand-alone components in view of baseband processors' distinct technical characteristics and function.

This is confirmed by the market investigation. For instance, one device OEM respondent explained that from the demand side "[w]hile the [application processor] and the baseband processor can be combined into a single integrated [System on Chip], the two components perform very different functions. The [application processor] cannot be used for cellular communication, and the baseband processor cannot be used to control and run the multitude of application programs and functionalities in today’s complicated smartphones". From the supply side, that same respondent indicated that "[t]he [application processor] and baseband are on very different technology and process roadmaps."

During the phase II market investigation, a majority of responding device OEMs confirm that the standalone and integrated solutions are viable alternatives for the provision of baseband and application processor functionalities in mobile devices.

Results were more varied when device OEMs considered the feasibility of replacing a discrete solution with an integrated solution (or vice versa) from a technical and commercial perspective, with respondents however indicating that technically such replacement would be possible during the design phase of a product.

The value of the mobile end-product does not appear a determining factor for whether one or the other architecture is chosen; in other words high-end mobiles can adopt either a standalone (AP+BC) or integrated (AP/BC) architecture.

Integrated baseband chipsets are in fact used in the majority of mobile devices. The most important buyer of standalone chipsets is Apple that combines its own custom-made application processor with a third-party baseband chipset.

c) Segmentation by cellular standard

The Commission considers, in line with the results of the market investigation, that baseband chipsets should be segmented by cellular standard.

For a mobile device to communicate with a cellular network, the device must be equipped with a baseband chipset that complies with the cellular standard supported by that cellular network. Baseband chipsets of an earlier generation will not support a later generation cellular standard. Unless baseband chipsets are backward compatible (multi-mode chipsets), the later generation baseband chipset will also not work on networks supporting earlier generation cellular standards.

As explained by Samsung, the various cellular standards are independent technologies not compatible with one another. Therefore, for a chipset to work on several networks, it must be multi-mode.

In the European Economic Area ("EEA"), the second generation GSM, the third generation UMTS and the fourth generation LTE represent the most wide-spread cellular technologies and are used by MNOs in parallel. The CDMA standard, which is widely deployed in the US and China, is not used in Europe. More than [50-60]% of all new mobile phones sold in Europe now support LTE. On a global level, chipsets supporting LTE are forecasted to reach [50-60]%, multi-mode UMTS chipsets [20-30]%, and GSM chipsets [20-30]% of industry shipments of baseband processors in 2017. By 2020, the chipsets supporting LTE will account for [60-70]% of basebands processors shipped.

Earlier generation baseband chipsets do not exert competitive constraints on baseband chipsets supporting a later generation standard; the various cellular standards are not substitutable either from a demand or supply side perspective. More specifically, multi-mode LTE chipsets which are backwards compatible with earlier generation cellular standards still in use (namely UMTS and GSM) are not substitutable with single-mode GSM chipsets, multi-mode UMTS chipsets, or single-mode LTE chipsets.

The market investigation confirms the lack of substitutability between standards. MNO respondents explain the need to continue supporting various cellular standards. Telenor, for example, explains that while "[i]n some EEA countries where Telenor operates, for instance in Denmark and Norway, UMTS and GSM is expected to be phased out and replaced with 4G and 5G only. However, due to a long transition period (years) and to support roaming networks with mainly GSM and UMTS technologies, multimode terminals from phone-manufacturers will be required for many years ahead".

The data rates achieved by earlier generation cellular baseband chipsets, in particular GSM but also UMTS, are insufficient for today's smartphones, the applications running on them, and do not meet the expectations of smartphone users. LTE technology allows mobile devices, with an LTE baseband chipset, to receive and transmit large amounts of data over a LTE cellular network. As explained by Intel, chipsets supporting only 2G or 3G standards are not suitable to transfer such data volumes and are therefore not substitutes for LTE chipsets.

Device OEMs therefore require chipsets supporting a specific (or several specific) standard(s), and may have limited demand for other standards, for example CDMA in view of its limited geographic deployment (including its absence in the EEA). In addition, according to the Linley Group, by now among smartphones "[…]".

Single-mode LTE chipsets are also not substitutable with earlier generation chipsets given that LTE coverage is still not ubiquitous and that voice communication is not supported by LTE mobile networks, which tend to be used for the transfer of data. Although single-mode LTE cellular networks using LTE technology for both data and voice (namely "Voice over LTE" or "VoLTE" technology) exist, the Commission understands that MNOs mostly provide multi-mode network service on their limited bands; the LTE network is used for transferring data while GSM, UMTS and CDMA networks are used for voice communication, but also data transfer where LTE network coverage is not available.

MNO respondents to the market investigation confirm the need for multi-mode cellular coverage. Among MNO respondents, most respondents do not see single-mode LTE chipsets as substitutable with earlier generation or multi-mode LTE chipsets. Among those indicating substitutability, Iliad explains that they are substitutable only "[f]or some (very limited) applications, like internet of things or datacom devices" while Tele2 considers that with the introduction of VoLTE, LTE has "a possibility to give the same coverage and voice services as GSM and UMTS and by then there is no need for GSM and UMTS".

Suppliers of a baseband chipsets complying with specific cellular standards (especially earlier generation standards) are unable to switch to the supply of more recent cellular standards, in particular LTE, in a short timeframe and without incurring significant additional investments (in costs and time) or risks.

The difficulty of switching is confirmed by the market investigation: most respondents consider it either difficult or very difficult for suppliers of single-mode GSM chipsets, single-mode LTE chipsets, or multi-mode chipsets compliant with both GSM and UMTS (but not LTE), to switch to the supply of multi-mode LTE chipsets.

See responses to questions 4-7 of Q9 – Questionnaire to device OEMs; responses to question 7 of Q13 – Questionnaire to Mobile Network Operators.

See The Linley Group: "Mobile Semiconductor Market Share Forecast 2015-2020", pp. 16-17; Form CO, Annex 4.16 [Doc ID: 327].

Intel, in response to the market investigation, indicates that "[a]n LTE-only [baseband chipset] is not substitutable for a multi-mode [baseband chipset] because it does not offer backward compatibility and because it cannot be used for voice communications. Backward compatibility is needed because most carriers still have areas in their network that have not been upgraded to LTE. The voice issue arises because carriers have not yet implemented the VoLTE standard for enabling voice communications over the LTE networks", see response by Intel to question 2.3.1 of Q10 – Questionnaire to baseband competitors, [Doc ID: 2307]; Pegatron explains that "the chipset compliant with LTE standards only cannot support the functionality that end users need nowadays, e.g. in the place where LTE signal coverage is limited, it is essential for the mobile to switch UMTS mode in order to make/receive calls.", see response by Pegatron to question 2.3 of Q9- Questionnaire to device OEMs, [Doc ID: 2054].

Switching from the supply of a single-mode LTE chipsets to multi-mode LTE chipset would also not be easy, and as one competing chipset supplier explains: "it is likely that a supplier of single-mode LTE chipsets would have to invest significantly in both design and development costs, tangible assets in the form of measurement system and research and development equipment, testing costs, and additionally required certifications in order to switch from supplying an LTE-only chipset to a multimode LTE chipset".

One device OEM respondent also indicates that not all suppliers have the technological know-how or required licences to produce chipsets compliant with the various technologies, giving Qualcomm's LTE chipsets compatible with CDMA as an example where competing suppliers of baseband chipsets may be unable to offer an alternative chipset supporting both LTE and CDMA. It is argued that this is the case because "Qualcomm does not provide licenses to its proprietary CDMA SEPs to chip manufacturers, and Qualcomm is by far the largest owner of CD[MA] SEPs".

d) Merchant v captive sales

For the purposes of this Decision, the Commission considers that captive sales of baseband chipsets should be excluded from the relevant product market as it is unlikely that such captive capacity could effectively constrain the merchant market.

Although the Notifying Party argues that device OEMs' internal production of baseband chipsets should be included within the relevant product market, the Commission notes that only certain device OEMs (in particular Samsung and Huawei) have internal production capacity and appear to use that production exclusively or almost exclusively for internal purposes and not to supply third party device OEMs. While internal production of baseband chipsets allow an OEM to partially limit its dependence on third party baseband chipsets, the main device OEMs with in-house production, namely Samsung and Huawei, still source a significant part of their supply of baseband chipsets from the Notifying Party.

Samsung has indicated the existence of significant entry barriers and obstacles for a device OEM already manufacturing baseband chipsets for its internal demand to start selling them on the "merchant market" for baseband chipsets. In explaining how "captive sales" by a device OEM influences the competitive parameters on the "merchant market" for baseband chipsets, Mediatek refers to captive sales being considered a "separate market from merchant sales".

e) Conclusion

The Commission considers that the relevant product market consists of both standalone and integrated baseband chipsets, segmented by cellular standard. In particular, chipsets compliant with LTE are not constrained by chipsets compliant with other cellular and non-cellular connectivity technologies. In addition, single-mode LTE baseband chipsets do not exert a constraint on multi-mode LTE chipsets that are also compliant with UMTS and GSM. In the following, unless specified, "LTE (baseband) chipsets" refers to multi-mode baseband chipsets that are compliant with LTE, UMTS and GSM. LTE baseband chipsets may or may not be compliant also with the CDMA cellular technology. Non-cellular wireless connectivity standards such as Wi-Fi and WiMAX do not constitute a competitive constraint on cellular baseband chipsets and are therefore excluded from the relevant product market. Finally, captive production of baseband chipsets, are excluded as captive capacity is not likely to effectively constrain the merchant market.

5.2.3.2. Near Field Communication/Secure Element technology

NFC technology is a short-range wireless connectivity standard composed of circuitry and software that uses magnetic field induction to enable communications/exchange of data between devices when such devices touch or are brought within a few centimetres of each other. NFC generally operates in a frequency range of 13.56MHz +/- 7kHz and at a distance of ten centimetres or less.

NFC chips are radio chips that support the NFC connectivity standard and enable devices to have: (i) a read/write mode (where one NFC-enabled device, interacting with another NFC-enabled device, can read in the second device or write data out to it); (ii) a card emulation mode, enabling the device to be used in a contactless infrastructure to perform applications such as payments, ticketing, access control, transit, and so forth; and (iii) a peer-to-peer mode, to exchange larger sets of data (files between smartphones, receive loyalty points, and so forth).

In order to ensure the security of NFC-based communications, NFC chips can be combined with various security technologies, in particular SE. SEs are tamper-resistant physical MCU chips that provide security and authentication in conjunction with NFC components to guarantee that data stored and data transmitted in an NFC communication are protected by an additional hardware-based layer of security. The SE stores payment information and authenticates transactions in a physical IC component. The SE microcontroller includes a secure operating system ("SE OS") that manages the functionality of the NFC MCU and runs the secure applications on the SE. The SE can be incorporated directly into a mobile device in the form of an "embedded" SE, or alternatively included on a mobile device's SIM card or Micro-SD card.

In an SE-secured NFC solution, there are three distinctive elements: (i) the NFC controller/chip; (ii) the SE; and (iii) the SE OS.

Such solutions on mobile devices can emulate multiple different types of protocols, including for instance EMV and others, and enable a mobile device to interact with a variety of readers that utilise different protocols, to perform functions such as payment and transit.

NXP is a leading supplier of NFC solutions, a technology it initially developed together with Sony. NXP holds a portfolio of NFC patents, both standard essential patents (SEPs) and non-SEPs.

NXP develops and supplies both standalone NFC controllers/chips and integrated NFC/SE combined solutions ("NFC/SE combined solutions"). NXP supplies three types of NFC chips, namely NFC frontends, NFC controllers, and NFC tags. Currently, NXP supplies only embedded SEs, integrated in the NFC/SE combined solution. NXP does not supply embedded SEs or SE OS as standalone products, but always in conjunction to an NXP NFC/SE combined solution. The Notifying Party explains that such combined solutions can either be in the form of a monolithic integrated circuit in a single die incorporating both the NFC and SE component or in the form of a NFC/SE combined solution chip included in a single package as a stacked die comprising two separate dice per technology. NXP currently supplies a stacked die combined solution with all components developed in-house (including NFC, SE and SE OS), and is developing a single die combined version of this chip to be available for commercial launch by end 2018.

Linked to its NFC technology, NXP has also developed MIFARE, a proprietary contactless security technology platform used in particular in transit ticketing/fare collection and similar applications (see Section 5.3.3). MIFARE can be used in connection with NFC on mobile devices, but also in other form factors, such as single-use passes, smartcards, ID badges, and fobs.

Qualcomm has recently exited the NFC product space. It does not develop or sell NFC or SE chips, but holds a number of NFC patents in its portfolio. Qualcomm stopped its internal NFC development programme in […] and ceased the shipping of its NFC chips in […]. Instead of continuing to develop its own NXP solutions, Qualcomm partnered in 2015 with NXP to include NXP's NFC chips in so-called "reference designs" for Qualcomm's baseband chipsets, thus providing a "showcase" or "instruction set" for device OEM customers to follow, so as to ensure seamless integration between Qualcomm's baseband chipsets and NXP's NFC chips.

A) NFC chips

The Notifying Party considers that the precise product market definition in relation to NFC chips can be left open as the Transaction would not raise any competition concerns.

The Notifying Party explains, however, that NFC is not the only secure mobile payment form but various alternative technologies exist that can perform similar functionalities to NFC chips. Such alternatives include Bluetooth Low Energy, Quick Response codes, and Samsung's proprietary Magnetic.

Devices with MIFARE embedded will only work with MIFARE enabled reader infrastructure. On mobile devices, it can be installed via replaceable SIM cards, embedded Universal Integrated Circuit Cards ("UICCs") and embedded SE chips.

The Notifying Party explains that BTLE is already used as an alternative technology to NFC, for example in the hospitality sector for the purposes of hotel room access, but is also an option for mobile payments in the retail sector. Compared to NFC, BTLE has a longer range, which the Notifying Party considers an advantage as it could "reduce friction during the checkout process" as consumers would not need to take their mobile device out and tap the reader to perform a mobile payment. NFC has been longer on the market than BTLE, which gives NFC an advantage, as this has resulted in more point of sales (PoS) terminals having NFC technology incorporated in view of EMV adoption. BTLE is, however, available on the majority of mobile phones, including the vast majority of iOS and Android devices (as well as emerging platforms).

The Notifying Party explains that QR codes are two-dimensional printed codes, similar to bar codes that interact with a smartphone or other mobile device through the camera (when the device is reading QR codes) or screen (when the device is generating readable QR codes). QR codes are being used widely for mobile payments and mobile transit applications and are garnering support from major banks that

Secure Transaction ("MST") technology, "HotKnot" technology, and Wireless Magnetic Communication ("WMC") technology.

The Notifying Party submits that the use of mobile devices to execute secure mobile transactions is still nascent, with the demand for products enabling such transactions still developing and with no single technology having become ubiquitous. Rather, there is inter-technological competition, with NFC and SE solutions facing competitive constraints from such other technologies. Different players, such as ecosystem developers, banks and retailers, mobile operators, and device OEMs all have an interest in influencing and shaping the way how mobile payment transactions should be executed and may opt for one or the other technological solution.

The Notifying Party argues in its reply to the Article 6(1)(c) decision that product characteristics such as the type of wireless connectivity, features, power consumption, perceived level of security, are important drivers of demand but functionality is what matters most, with the predominant purpose of NFC and SE chips the execution of mobile payments through mobile devices. To this end, there are, however, several competing technologies available, none of which has become predominant; any assumption that NFC and SE chips would become ubiquitous is therefore not grounded in facts.

Specifically as regards NFC, the Notifying Party claims that technologies such as BTLE, QR codes and MST all share the same end use, namely of executing mobile payments, and the core functionality of enabling the communication between the seller and purchaser in a mobile payment transaction. BTLE, QR codes and MST all exert competitive constraints on NFC solutions and should be seen as substitutes rather than complements to NFC technology. Characteristics may differ in terms of the power efficiency of the technology, or required proximity to perform a communication, but the core functionality remains the same. Device OEMs might have specific preferences as to characteristics, but individual preferences are not representative of a whole industry.

The Notifying Party explains that MST was invented by Samsung and the technology mimics the magnetic strips on credit cards, allowing mobile devices to be used anywhere a magnetic strip reader is available at PoS terminals. MST technology functions in card emulation mode to act like a contactless smartcard, mainly used for payment applications. MST is more broadly deployed than NFC but is currently only used in Samsung's proprietary Samsung Pay ecosystem, which uses a combination of technologies, namely NFC, HCE, QR codes and MST, giving Samsung Pay the flexibility to work with a number of different PoS infrastructure systems. Samsung Pay works with PoS terminals in places where NFC is not available, and because any magnetic credit card reader can be used, MST is actually is more versatile than NFC. The Notifying Party concedes that the deployment of MST in the EEA is limited to four banks in Spain. The fact that the majority of PoS infrastructure in the EEA is chip and PIN based, rather than magnetic stripe-enabled PoS, poses a challenge for its success in the EEA.

The Notifying Party explains that HotKnot, a proprietary technology of MediaTek, using proximity and gravity sensors in conjunction with a mobile device’s touchscreen. Devices featuring HotKnot use a touch sensor chip to send communication protocols, gravity sensor to ensure the actual contact between devices, and a proximity sensor to determine whether the two electronic devices are close enough to one another. As an advantage to NFC, HotKnot does not require an antenna or RF communication/NFC chip for data to be transferred, making it cheaper and easier to deploy in smartphones. The technology is used in China and included on OPPO, Joy and Yoyo smartphones.

The Notifying Party explains that WMC is a payment technology by LG and Dynamics Inc. which is similar to Samsung's MST, namely it allows for over-the-air transmission of data to PoS terminals with magnetic stripe readers.

The Notifying Party explains that the use of mobile devices to execute secure mobile transactions is still nascent, with the demand for products enabling such transactions still developing and with no single technology having become ubiquitous. Rather, there is inter-technological competition, with NFC and SE solutions facing competitive constraints from such other technologies. Different players, such as ecosystem developers, banks and retailers, mobile operators, and device OEMs all have an interest in influencing and shaping the way how mobile payment transactions should be executed and may opt for one or the other technological solution.

The Commission considers that NFC chips (and the underlying technology) constitute a separate product market, distinct from other technologies which can enable mobile payment and transport ticketing/fare collection transactions, in particular BTLE, QR codes, and MST.

The Commission has previously assessed NFC based payment systems in the context of mobile payment platforms but left open the exact market definition for payment systems, including any solutions involving NFC and SE chips.

The market investigation confirms that NFC chips are distinguishable from other technologies such as BTLE, QR codes and MST based on product characteristics, price and intended use or application.

The market investigation also confirms that while other technologies besides NFC exist (in particular BTLE, QR codes, and MST) for the specific function of enabling a mobile payment/transport ticketing transaction, these are not viable alternatives. Such technologies have their distinct traits and are adopted by (or available to) device OEMs to varying degrees. There are also specific regional disparities in their adoption rate (for example, wide-spread use of QR codes for mobile payments in China, but insignificant deployment elsewhere).

Although various technological solutions could be used to enable a mobile payment transaction, there are currently, in practice, no viable alternatives to NFC. As Gemalto explains, NFC, BTLE, QR codes and MST are all proximity contactless technologies, but NFC technology used for the purposes of mobile payment is the "only fully backward compatible technology that enables users of mobile devices to benefit from existing services and offer a seamless user experience, with no compromise or restriction of service".

NFC is also required for mobile transit ticketing solutions based on MIFARE, which is the leading transit service technology today. STMicroelectronics explains that MIFARE can "only be addressed through NFC technology from a phone or a device", Gemalto explains that "if a device only had an eSE interfaced with a B[T]LE chipset, the device could not support services such as MIFARE that require a secure NFC environment", whereas Giesecke & Devrient ("G&D") notes that currently, it is a requirement that "ISO/IEC 14443 Type A standard is used as communication protocol, therefore it is not possible to implement MIFARE using different communication protocol other than NFC".

There is a distinct demand by device OEMs for NFC based solutions for mobile devices. NFC is quickly becoming a standard feature in most smartphones. According to market forecasts by the Linley Group, [80-90]% of all smartphones and [60-70]% of all handsets (up from [30-40]% in 2015) are expected to include NFC by 2020. The Linley group predicts that "[…]".

Market players perceive other technologies (in particular BTLE, QR codes and MST) as complements to NFC, rather than substitutes. Such technologies are not perceived as viable alternatives by most responding NFC competitors and baseband chipset suppliers, or by the main device OEMs.

On the customer side, the largest device OEMs agree that NFC technology in smartphones is becoming a standard feature which is ultimately required by the market players further downstream, in particular network operators. In light of the requirements imposed by downstream players, device OEMs cannot switch away from including NFC chips in their mobile devices. Samsung explains that "[a]s long as Samsung's customers require NFC and SE, Samsung Mobile cannot substitute other technologies for those components. It is possible, where requested, to add any or all of the above-mentioned technologies; however, they are not alternatives to NFC/SE". LG Electronics considers that the technology strategies of "tier 1 OEMs" such as Apple and Samsung has an influence on the "trend for mobile payment and/or fare collection" and in that respect "BTLE, QR Codes and MST do not yet constitute viable alternatives to NFC".

Some responding device OEMs indicate that while the various technologies (NFC, BTLE, QR codes, and MST) can all be used to perform mobile payment and fare collection transactions, not all have the same "traction" in the payment services industry, and have thus not reached the same adoption rate as NFC.

a) Substitutability between NFC and BTLE

Compared to BTLE and other radio technologies, NFC with its limited range of a few centimetres, is […] than other technologies.

The weakness of BTLE for the purposes of mobile payments and transport applications is also confirmed by NXP's internal documents. NXP lifts up, in particular, the lack of infrastructure, security concerns, and weak user experience of BTLE for mobile payment and transit use cases.

Contrary to NFC, BTLE is not perceived by the market as a well-suited technology for mobile payment transactions and lacks the security that NFC solutions allow for. BTLE is better adapted for other types of use cases, for example contactless speakers, earphones, car connection. Gemalto explains that BTLE, due to its "functional distance range", does not meet the requirement in payment and transit ticketing transactions for the individual user to explicitly endorse a transaction (for example by tapping the mobile to the POS reader). Infineon concurs, and explains how NFC was designed with security in mind and requires close proximity with the reader terminal to function, which constitutes a "clear advantage" of NFC over BTLE for the specific purposes of mobile payments and ticketing transactions. Infineon notes that "bringing the device into the proximity of the terminal is considered both an element of security to ensure that only the device that is very close to the terminal is used for payment as well as an expression of user intent to pay. BTLE in the contrary is defined for longer distances between any two devices and would not meet those requirements". Intel argues that due to the security and usability shortcomings of BTLE, it is a "poor substitute for NFC" for the purposes of mobile payment and fare collection. Trustonic points out that BTLE being a long range technology, its usability for mobile payments might, in practice, create difficulties for the POS terminal to "select the right nearby user for the payment transaction". Samsung explains that BTLE, contrary to NFC, requires that the user manually sets up the connections between devices, a process which is likely to take more time compared to the almost instantaneous connection via NFC. Apart from being quicker, NFC also provides for more security, also in situations when there are multiple devices in within close proximity. Another device OEM points out that "[f]or mobile payments there are no real alternatives to NFC/SE" with BTLE "a solution better adapted for functionalities other than mobile payment transactions".

BTLE based mobile payment solutions cannot also benefit from the existing POS infrastructure in place which are increasingly NFC enabled, and the wider deployment of BTLE for mobile payment uses would necessitate important investments in such infrastructure.

BTLE and NFC may have, though, complementary usages.

b) Substitutability between NFC and QR codes

Similarly to BTLE, QR codes are also not equally well suited for enabling mobile payment transactions as NFC. Although QR codes are widely used for mobile payment in China (but also India), their use elsewhere is limited. The solution is deemed to suffer from low security. EMVCo, the technical body managing the EMV payment specifications, has recently, however, issued specifications also for QR code payments with its efforts "focused on creating QR Code payment specifications that provide convenience, security and reliability in line with other EMV Specifications".

The weaknesses of a QR code based mobile payment and transport ticketing solution is confirmed not only by market participants responding to the Commission's market investigation but also by NXP's internal documents. In a presentation of 2016, NXP lifts up, in particular, the security concerns of QR codes for mobile payment and transit use cases, the limited infrastructure in place and the weak user experience that QR codes offer in particular in transport applications. In the market investigation, although some respondents saw QR codes and NFC as substitutes, others considered them as complementary to each other.

Among the main market participants, the weaknesses of QR codes compared to NFC is confirmed by competing players such as Infineon, STMicroelectronics, and Gemalto. One device OEM points to the very low security level that QR codes provide compared to NFC/SE, highlighting that while QR Codes are used in China for mobile payments, their use is limited to online payments with limitations also as to the number and size of daily/monthly transactions.

Gemalto explains that QR codes, by its "optical reading nature" and the consequent weakness in "reading accuracy" and "user experience", are also badly adapted to ticketing and fare collection transactions that require high performance solutions to enable a "seamless flow of passengers". To function, the mobile device also needs to be powered up (it does not function in a battery-off mode) and online.

c) Substitutability between NFC and MST

Although developed for enabling mobile payment transactions, MST technology is not a viable alternative to NFC, but rather a temporary solution, specific and limited to Samsung, to allow for backward compatibility with magnetic stripe card technology and infrastructure. Magnetic stripe technology used for payment transactions has increasingly been replaced by the EMV standard/chip card technology, especially in Europe and the Single Euro Payments Area ("SEPA"), thus limiting the relevance of MST in the EEA.

MST is a proprietary Samsung technology, "exclusive to Samsung", which is not made available to third parties, including other device OEMs. Samsung explains that while MST allows Samsung (and Samsung Pay) to connect with legacy (non-NFC) POS terminals with a magnetic stripe function, demand for MST is expected to decrease as more NFC-enabled POS terminals become available. Samsung expects this number to reach above 90% as a result of the requirement by Mastercard for POS terminals in Europe to support NFC by 2020.

Samsung explains that MST is intended as a "means to quickly penetrate the market by using existing magnetic card readers that are available at almost all retailers, and is not intended to replace NFC, because most financial institutions have already established roadmaps for NFC-based payment. Some countries prohibit using MST for making payment. If Samsung adopts [a] new type of technology instead of NFC, then not only Samsung but also all stakeholders in the payment system need to replace their infrastructure (e.g., merchant POS terminals, acquirers'/issuers' spec changes, networks' spec change, service provider app changes).

d) Installed base of POS terminals

A key element differentiating NFC payment solutions from other available technologies for the enablement of a mobile payment transaction is the widely deployed installed base of NFC-enabled POS terminals within the EEA.

Gemalto, a supplier of SE OS, notes that only NFC/SE solutions offer backward compatibility with existing commercial services for contactless cards (payment, ticketing and identity), and thus constitute "natural technologies to ensure business continuity for these legacy existing services when migrating from cards to mobile" and benefit from the "full ecosystem of readers" already deployed. The existing reader infrastructure constitutes a significant investment, and switching away from it to equipment based on an alternative technology would be very costly. Gemalto points out that NFC proximity connectivity is the only option that can benefit from the legacy ISO/SEC 14443 reader installed base. STMicroelectronics, a supplier of NFC and SE chips, explains that the ability to "immediately comply with the existing Payment/Transport infrastructure deployed all over the world" holds true to mobile devices integrating NFC but does not apply to other technologies, for example BTLE or QR codes. The existing POS infrastructure in place in the EEA does not support such technologies.

Interoperability of POS terminals and NFC is ensured through standardisation (ISO-14443 and NFC Forum specifications, also based on ISO-14443). Infineon explains that contrary to NFC enabled POS which exist already, there is no "available infrastructure to support" BTLE, MST and QR codes in the EEA.

According to analyst firm Berg Insight, NFC-enabled points of sale terminals continue to show strong momentum in 2016 with more than 90% of new POS terminals shipped in 2016 in Europe and North America NFC-ready. Furthermore, on the global level the installed base of NFC-ready POS terminals is expected to grow from 45 million units in 2016 to 86.9 million units in 2020.

e) Merchant v captive sales

For the purposes of this Decision, the Commission considers that captive sales of NFC chips should be excluded from the relevant product market as it is unlikely that such captive capacity can effectively constrain the merchant market beyond the ability of device OEMs with internal production to limit their reliance on third-party suppliers.

The Commission notes that among device OEMs only Samsung has internal production capacity and appears to use that production exclusively or almost exclusively for internal purposes and not to supply third party device OEMs. While this internal production allows Samsung to partially limit its dependence on third party suppliers of NFC chips, the mobile device OEM still sources part of its supply of NFC chips on the merchant market from NXP.

Captive sales do not constitute a competitive constraint on the merchant market, except to the extent that those device OEMs with in-house supply of NFC chips are able to limit their reliance on third party suppliers for those specific components.

f) Conclusion

The Commission considers that other technologies which are available to enable a mobile payment or transit ticketing transaction such as BTLE, QR codes and MST are distinct from NFC and do not exert significant competitive pressure over NFC. NFC chips (including the underlying technology) should therefore be considered as a separate product market. Only NFC chips sold on the merchant market should be included in the relevant product market.

B) Secure Element

1. Notifying Party’s views

The Notifying Party considers that the precise product market definition for the SE functionality can be left open as the Transaction would not raise any competition concerns.

The Notifying Party argues, nonetheless, that there are alternative security technologies to the SE, both hardware and software based, that provide for secure authentication of mobile transactions and can facilitate mobile payment and other applications. SEs face inter-technological competition from a variety of technologies, including Host Card Emulation ("HCE"), Trusted Execution Environment ("TEE"), and Trusted Platform Modules ("TPM"), the most popular solution for mobile payment authentication currently being HCE. The Notifying Party argues that many of those technologies will co-exist in the future, and that the choice of technology by the device OEMs will depend on various factors including security requirements, regional specificities, user experience, extent of use cases supported, and legacy infrastructure.

https://www.nfcworld.com/2016/11/16/348537/nfc-ready-pos-terminals-to-hit-16m-units-worldwide-in-2016/, [Doc ID 3294]

See in particular The Linley Group: "Mobile Semiconductor Market Share Forecast 2015-2020", page 84, Form CO, Annex 4.16 [Doc ID: 327]; non-confidential responses by Samsung to questions 8 and 15 of RFI 28, [Doc ID: 2671].

The Notifying Party explains that there are different types of attach for SE chips (eSE, which stands for embedded SE in the smart mobile device and "UICC-based SE" Universal Integrated Circuit Card where the secure element resides in the SIM card).

The Notifying Party explains that HCE is a software-based solution which is compatible with the same PoS infrastructure as the SE. HCE is the most popular alternative to SE, and is employed in the Android Pay platform. HCE is a viable alternative to SE to carry out secure transactions. The Notifying Party explains that from a functionality perspective, both solutions can be applied to the same end-uses (including mobile payment, transport), although the two technologies function in their own manner,

150

151

("TEE"), and Trusted Platform Modules ("TPM"), the most popular solution for mobile payment authentication currently being HCE. The Notifying Party argues that many of those technologies will co-exist in the future, and that the choice of technology by the device OEMs will depend on various factors including security requirements, regional specificities, user experience, extent of use cases supported, and legacy infrastructure.

The Notifying Party argues that those alternative technologies exert competitive pressure over SEs. The Notifying Party refers to a number of device OEMs (namely Huawei, Sony, LG) already using software-based security solutions instead of an SE chip and argues that mobile operating system suppliers (such as Apple, Microsoft) can similarly to Google (with its Android Pay) also implement a software solution on their mobile operating system instead of or in addition to a physical SE.

As for the SE OS, the Notifying Party explains that the OS is always sold in combination with the secure hardware (for example an SE chip), with the integration of the software with the hardware taking place before shipping to the device OEM. NXP supplies the hardware and software (namely NXP's SE OS) components of its SE […]. NXP also […].

Competing SE providers that do not have an SE OS developed in-house, source the SE OS from a third party supplier, such as Gemalto, Oberthur and G&D. It may therefore be possible, according to the Notifying Party, to distinguish a narrow market for the supply or licensing, or both, of SE OS to SE chip suppliers; there is no distinct demand for stand-alone SE OS from device OEMs.

In response to the Article 6(1)(c) decision, the Notifying Party argues that the Commission's reasoning and the results of the phase I market investigation do not provide support for the claim that SE chips constitute a separate relevant product market. Rather, the Notifying Party puts forward arguments in support of a market definition encompassing at least HCE and SE. The Notifying Party claims, in

different from the other. Performance wise, HCE and SE solutions are comparable but differ to the extent that a SE based solution, through its dedicated MPU, offers slightly faster transaction speeds, is always available, and requires less processing capability. The cloud-based HCE requires network connectivity, occupies more memory, and is more demanding from a processing perspective. But HCE is scalable and does not use additional physical space in the mobile device, which is an advantage in small/slim mobile devices. For OS vendors, HCE ensures that customers (device OEMs) can be offered the same payment security solution irrespective of the customers' hardware.

150 The Notifying Party explains that TEE is a secure area of the main processor in a smartphone (or any connected device) that ensures that sensitive data is stored, processed and protected in an isolated, trusted environment. TEE enables end-to-end security by enforcing protected execution of authenticated code, confidentiality, authenticity, privacy, system integrity and data access rights. Compared to other security environments on the device (like SE, HCE and TPM), TEE also offers high processing speeds and a large amount of accessible memory. TEE technology is used in mobile payments to protect payment credentials while a transaction is being authorised and offers a trusted user interface which ensures that the correct information is displayed to the user and that the information displayed on screen and entered by the user is secure. TEE technology can also be used for mobile public transport functions. TEE is a “tamper-resistant” function block in an AP, which can be programmed to ensure that sensitive data is stored, processed and protected in an isolated trusted environment. TEE resists against all software attacks as well as the physical attacks performed on the main memory of the system. TEE can be used in conjunction with HCE, in which case the HCE software is installed in the TEE.

151 The Notifying Party explains that TPM is a hardware based solution, namely a chip that provides security functions, found most commonly as a component on motherboards of laptops and desktops for corporate or government users. They can also be installed in consumer devices and are sold as independent components.

The Notifying Party explains that HCE as an alternative technology to SE cannot be ignored and refers to the "strength of its proponents", mentioning Google and Microsoft and to its wide use across mobile devices. The Notifying Party argues that HCE is supported as a security technology for mobile payment transactions by credit card providers such as Mastercard and used by Android Pay when executing a payment, "even if that particular device includes an SE". While HCE may provide a lower security level than SE, such level is, according to the Notifying Party, "a choice, not a requirement, for mobile payments" and not a sufficiently distinctive element to distinguish SE chips from other types of security technology solutions. Also, how different security technologies affect the speed of a payment transaction or whether the HCE technology requires specific software or cloud-based architecture only reflect the fact that HCE and SE are alternative technologies. Further, the Notifying Party argues that while HCE can be applied cumulatively with a SE it does not signify that the two technologies cannot be seen as alternatives to each other. Finally, the Notifying Party reaffirms its claim that the secure element space remains "quite clearly nascent".

The Commission considers that while NFC mobile transactions can be secured through the use of an SE, which currently allows for the highest level of securitisation of a mobile transaction, there are also other technologies available which may be (and are) used for the same purpose. Those other technologies include, in particular, HCE and TEE.

In the Article 6(1)(c) decision the Commission took the preliminary view that SE chips, including the SE OS, would constitute a separate product market, distinct from HCE, TEE and TPM, and consequently carried out its competitive assessment on that basis. Following the phase II market investigation, the Commission considers that the product market for securing NFC based mobile payment transactions may, however, be wider than simply SE chips (including the SE OS), and include other security technologies such as in particular HCE and TEE.

For the purposes of this Decision, the exact product market definition can be left open.

The phase I market investigation indicated that SE chips should be distinguished from other types of security technologies used for payment and transport applications such as HCE, TEE, and TPM based on product characteristics, price, intended use or application, but responses were mixed when asked whether such technologies would be complementary or alternative to SE. A mixed picture emerges also following the phase II market investigation. In particular HCE and TEE are mentioned by respondents as viable alternatives to SE in securing an NFC transaction. TPM, on the other hand, is seen as a technology developed for the PC world, and targeting use cases different from mobile payment and ticketing; it is not deemed an alternative technology to secure NFC transactions; one respondent sees it as a "function" of the embedded SE.

The main factors distinguishing SEs from other technologies are the high level of security and speed that SEs can provide to payment and ticketing transactions. On the other hand, technologies such as HCE, TEE are "easy to deploy and less cost[ly], but [provide] less security", and in addition do not work without battery (HCE and TEE) or offline (HCE).

Overall, however, and despite the numerous claims of the technologies offering a lower level of security than SE, the responses to the market investigation give a certain degree of support to the notion that alternative technologies, in particular TEE and HCE, may nonetheless provide a sufficient level of security for the purposes of mobile payment or fare collection transactions (or both) to constitute a competitive alternative to SEs. A certain degree of substitutability may, therefore, exist between the various technologies, especially from a demand side.

A distinguishing factor mentioned by market participants, which may, however, put SEs completely apart from other security technologies, is the proprietary NXP transit technology platform MIFARE, which is the leading transit service technology in the world today, and which currently only runs on the SE. To the extent that MIFARE is a required functionality on a mobile device, SEs would not be substitutable with any of the other security technologies for the purposes of NFC mobile transactions.

Supporting the notion that the market is wider than simply SEs, is the fact that mobile device OEMs already today employ various security technology solutions to secure NFC transaction on their devices. Existing mobile payment solutions use alternative solutions: Android Pay uses HCE or a combination of HCE and TEE, while Samsung Pay and Apple Pay rely on SEs for securing the NFC transaction. Samsung Pay can also use HCE in combination with TEE; the choice of technology being dependent on network carriers. Samsung explains as follows "[s]ome carriers (generally those with their own payment solutions which use the SE) prohibit Samsung Pay from using the SE, so in those cases, Samsung will be required to use TEE/HCE instead for Samsung Pay".

In its internal documents, NXP lists "[…]" for HCE technology, in particular its suitability for the purposes of securing mobile payments, ticketing/fare collection and access.

According an internal presentation of 7 April 2016, NXP explains that "[…]" and indicates that it provides a "[…]". According to NXP, the limitations that HCE

See for instance non-confidential response by Gemalto to question 1.1b) of RFI 25, [Doc ID: 2152]; non-confidential response by Infineon to question 1.1b) of RFI 29, [Doc ID: 2955]; non-confidential response by STMicroelectronics to question 1.1b) of RFI 30, [Doc ID: 2811].

See responses to questions 19, 20 of Q9 – Questionnaire to device OEMs; responses to questions 14, 15 of Q10 – Questionnaire to baseband competitors; responses to questions 9, 10 of Q11 – Questionnaire to NFC competitors.

See for instance non-confidential response by ZTE to question 20 of Q9 – Questionnaire to device OEMs, [Doc ID: 2147]; non-confidential response by Gemalto to question 1.1 of RFI 25, [Doc ID: 2152].

Most respondents are not able to provide an answer to this question.

See responses to question 22 of Q9 – Questionnaire to device OEMs; responses to question 17 of Q10 – Questionnaire to baseband competitors; responses to question 12 of Q11 – Questionnaire to NFC competitors.

suffers from do not apply to SE technology: from a security perspective HCE is deemed to pose a financial risk and raises privacy concerns; from an infrastructure scalability perspective, HCE is weak for various reasons including its lack of support of offline POS readers, lack of MIFARE support; and from a user experience perspective, HCE is only suitable for low value payments, low value tickets (single trip tickets), and low security access. As a key limitation compared to SE, NXP highlights the fact that HCE cannot be used for mobile payment or ticketing if the mobile device is without battery or offline.

In another internal presentation of June 2016, in the context of […], NXP lists the security and infrastructure limitations of HCE, claiming that […]. It also highlights the limitations of HCE for transit applications, namely […] .

The distinctiveness and superiority of SE compared to the other security technologies to be used with NFC is echoed by the main competitors in the NFC/SE space and customers/device OEMs. Infineon, a supplier of embedded SEs, considers that NFC implementation based on SE will "become the de-facto standard", and technologies such as TEE and HCE provide a lower level of security.

STMicroelectronics, a supplier of both NFC and embedded SE chips, explains that for mobile payment, the embedded SE has a "real advantage in term of intrinsic security and convenience of use" compared to HCE/TEE. Likewise, in mobile ticketing, STMicroelectronics suggests that only the embedded SE can fulfil the security certification levels required by for example MIFARE. Gemalto, a supplier of SE OS, also confirms the distinctiveness of the NFC/SE combination for the purposes of mobile payment, especially with MIFARE, by stating that although NFC could be combined with TEE and HCE, this would result in the provision of "less secure services not suitable for all use cases (an in particular, not suitable for transit applications based on Mifare for instance)".

NXP internal document, “[…]”, 7 April 2016, slides 4-5 [Doc ID 1457-24348].

NXP internal document, “[…]”, June 2016, slide 3 [Doc ID 1457-24347].

See for instance non-confidential response by Infineon to question 1.1b) of RFI 29, [Doc ID: 2955]; non-confidential response by Gemalto to question 1.1b) of RFI 25, [Doc ID: 2152]; non-confidential response by G&D to question 1.1.b) of RFI 26, [Doc ID: 1932]; non-confidential response by G&D to question 1.1b) of RFI 26, [Doc ID: 1932]; non-confidential response by STMicroelectronics to question 1.1b) of RFI 30, [Doc ID: 2811]; non-confidential response by Samsung to questions 6, 10, 11 of RFI 28, [Doc ID: 2671]; response by Samsung to question 9 of Q1 – Questionnaire to device OEMs, [Doc ID: 1092].

Infineon considers that TEE based implementations "can be compromised and misused"; HCE is described as a "cloud-based alternative" to SE for mobile payment use cases, albeit with "lower level of security". Infineon explains that the payment tokens in a HCE solution are stored in the open memory of a mobile device, but can also be stored on an embedded SE. Although seen as a low-security alternative to SE for mobile payments, Infineon does not consider HCE a viable alternative for mobile transport/fare collection use cases, as it is unable to "meet the speed requirements of transport applications". TPM is described as unsuitable for mobile payment and fare collection use cases ("The features and functions of a TPM product cannot be used to implement mobile payment and mobile ticketing applications"); TPM is used to ensure the integrity of devices and ensuring/supporting data protection, for example protecting the pass-code of BitLocker hard disk drive encryption in personal computers; see non-confidential response by Infineon to question 1.1 of RFI 29, [Doc ID: 2955].

STMicroelectronics further deems TPM "definitely not suitable to support use cases brought by NFC technology (such as payment and transport)"; see non-confidential response by STMicroelectronics to question 1 of RFI 30, [Doc ID: 2811].

Gemalto indicates that a combined solution of TEE, HCE and tokenisation methods "offers a valid option" for mobile payments, but has its limitations to the extent that it provides less protection for off-line payments, it does not work when the mobile device is in a battery-off mode, and it "requires an upgrade of the transaction processing backend".

NXP internal document, “[…]”, 7 April 2016, slides 4-5 [Doc ID: 1457-24348].

NXP internal document, “[…]”, June 2016, slide 3 [Doc ID: 1457-24347].

affirms that TPM is not an alternative technology to embedded SEs for securing an NFC mobile transaction, while HCE and TEE would necessitate "additional provisions/investments in the infrastructure" compared to an NFC solution based on the embedded SE. It admits that there is currently a "mixed landscape of eSE/non-eSE based mobile payment solution".

The distinctiveness of the SE compared to other technologies like HCE, TEE and TPM is supported also on the demand-side by device OEMs.

Samsung does not see HCE, TEE and TPM as viable alternatives to SEs, and explains that from a demand-side perspective, cloud-based solutions such as HCE "will continue to be viewed as providing inadequate security for most mainstream applications". An NFC solution secured by HCE alone "is not sufficient" and would need to be complemented by other technologies, such as SE, for the purposes of high-security financial transactions. Compared to hardware based security solutions such as TEE and TPM, Samsung considers the SE to have "significant head start", and the significant hardware investments already made by device OEMs with respect to SE "will have a "lock-in" effect going forward, regardless of the features presented by competing technologies". In explaining why the TEE is rather a complement to the SE, Samsung notes that the TEE is designed as a "hidden/mini" OS that "boots simultaneously with Android and has priority over Android", thus allowing it to "take priority over input and then securely transmit that input to a secure area without the risk of any Android functions intercepting such input." Samsung notes that because of this priority over Android, the TEE "has been further developed by Samsung to act as a "tunnel" when a user inputs sensitive input so that TEE will securely tunnel that input straight into the SE", thus complementing the function of a SE.

[…] . […] .

Another device OEM also confirms that TEE and HCE are not viable alternatives to the hardware based SE solution, not least because of lower security levels: TEE is more a software environment than a hardware solution (like SE) and it is also not possible to store data on TEE; HCE has not been very successful in Europe, one of the reasons being the much lower level of security it offers compared to SE.

Similarly to NFC chips, there is a distinct demand by device OEMs for NFC based solutions combined with a SE. For device OEMs, there is little option to switch away from NFC solution which relies on the SE for its security. This is all the more the case when MIFARE is needed. As Samsung puts it: "As long as Samsung's customers require NFC and SE, Samsung Mobile cannot substitute other technologies for those components. It is possible, where requested, to add any or all of the above-mentioned technologies; however, they are not alternatives to NFC/SE".

See non-confidential response by G&D to questions 1.1 and 1.4 of RFI 26, [Doc ID: 1932].

See non-confidential response by Samsung to questions 6 and 10 of RFI 28, [Doc ID: 2671].

See response by Samsung to question 9 of Q1 – Questionnaire to device OEMs, [Doc ID: 1092].

See non-confidential response by Samsung to questions 10 of RFI 28, [Doc ID: 2671].

See non-confidential response by Samsung to questions 11 of RFI 28, [Doc ID: 2671]; non-confidential response by Samsung to question 9 of Q1 – Questionnaire to device OEMs, [Doc ID: 1092].

On the supply side, most competing suppliers of NFC/SE solutions confirm the distinct demand for NFC combined with SE. Gemalto explains there is no switching by device OEMs from NFC solutions relying on the SE: Apple and Apple Pay make use of NFC and SEs, while Android devices, even where other security technologies than SEs are installed, those manufacturers "also continue to install eSEs". In fact securing NFC mobile transactions with a SE "remain mandatory for usages such as mass transit ticketing […]". STMicroelectronics argues that "at this stage, [STMicroelectronics] see most of large OEMs strongly committed to the usage of embedded Secure Elements" for the purpose of enabling mobile payment services. Infineon also expects continued reliance by device OEMs on SEs for mobile payments, as switching away from SEs to HCE or TEE would imply that OEMs (namely Samsung and Apple) would decide to "degrade security of their solutions", and this is not seen as a likely prospect.

Finally, as regards the SE OS, the market investigation has indicated that the SE OS should be distinguished from other types of software running on a smartphone due to its specific function and the applicable certification requirement. While most respondents also thought that the SE OS could be distinguished as a separate product from the SE chip, device OEMs indicate that they purchase SE OS in combination with SE chips rather than on a stand-alone basis. This confirms the Notifying Party's view that no distinct demand from device OEMs exist for standalone SE OS. NXP, in any event, does not license or sell its SE OS on a standalone basis to third parties, and is not, therefore, active on any (narrow) market for standalone SE OS.

Merchant v captive sales

For the purposes of this Decision, the Commission considers that captive sales of SE chips should be excluded from the relevant product market as it is unlikely that such captive capacity can effectively constrain the merchant market.

(168) The Commission notes that among device OEMs only Samsung and, to a very limited extent, Huawei have internal production capacity and appear to use that production exclusively or almost exclusively for internal purposes and not to supply third party device OEMs. While this internal production allows Samsung and Huawei to partially limit their dependence on third party suppliers of SE chips, the mobile device OEMs still source part of their SE chips on the merchant market .

(169) Captive sales do not constitute a competitive constraint on the merchant market, except to the extent that those device OEMs with in-house supply of SE components are able to limit their reliance on third party suppliers for those specific components.

See non-confidential response by Samsung to question 6 of RFI 28, [Doc ID: 2671].

See non-confidential response by […] to question 13 of RFI 31, [Doc ID: 2633].

See non-confidential response by Gemalto to question 1.4 of RFI 25, [Doc ID: 2152].

See non-confidential response by STMicroelectronics to question 1.4 of RFI 30, [Doc ID: 2811].

See non-confidential response by Infineon to question 1.4 of RFI 29, [Doc ID: 2955].

See responses to question 11 of Q1 – Questionnaire to device OEMs; responses to question 14 of Q2 – Questionnaire to baseband competitors; responses to question 14 of Q3 – Questionnaire to NFC competitors.

See responses to question 13 of Q2 – Questionnaire to baseband competitors; responses to question 13 of Q3 – Questionnaire to NFC competitors.

See responses to question 12 of Q1 – Questionnaire to device OEMs.

See non-confidential responses by Samsung to questions 8 and 15 of RFI 28, [Doc ID: 2671].

b) Conclusion

(170) The Commission considers that while there are a number of elements that support distinguishing SEs (including the SE OS) from other technologies, in particular HCE and TEE, the question as to whether the relevant product market should be wider than SEs to also include other technologies can be left open for the purposes of this Decision. The competitive assessment will be carried out on the possible product market of embedded SEs (including the SE OS), which is the market where NXP holds the most market power. Embedded SE is the only type of SEs currently supplied by NXP, and it supplies them only as part of the company's combined NFC/SE solution, and not as a standalone product. Carrying out the competitive assessment by using a wider product market definition, namely one which would comprise not only SE but also alternative technologies, in particular HCE and TEE, would be unlikely to raise any additional concerns, considering in particular the more limited market position of NXP on such wider product market .

C) Combined NFC/SE solutions

(171) The Commission considers that the NFC/SE combined solution constitutes a distinct product market, separate from the market for standalone NFC chips and SE chips (including SE OS).

(172) The market investigation has shown that device OEMs have different procurement strategies with respect to the components required for the NFC solution included on their mobile devices. To the extent that the device OEM opts for an NFC solution secured with a SE (instead of for example a combination of NFC with HCE and TEE), it can either purchase standalone components or a NFC/SE combined solution .

(173) Device OEMs tend to opt for the combined solution, for which there is a distinct demand from device OEMs.

(174) Samsung explains that depending on use-case, device OEMs may opt to procure only standalone NFC chips (especially when an application does not require a SE), while in other cases, device OEMs decide to purchase either a combined NFC/SE solution, or a standalone NFC solution for some of its smartphones and a combined solution for other smartphones, especially those smartphones implementing a mobile payment system such as Samsung Pay or Apple Pay .

(175) G&D points out that while various approaches are used by device OEMs to obtain "best of breed solutions" which correspond to their respective preferences "the continuous integration of the NFC and SE chips into a single package (SIP – System in Package) and even a single die (piece of silicon)" has led to the reduction of device OEMs' ability to combine standalone components by themselves .

See Form CO, paragraph 767; Annex 4.1 to the Form CO, paragraphs 156, 161.

In relation to alternative technologies such as TEE, HCE and others, the Notifying Party indicates in the Form CO, paragraph 822 "To the best of the Parties’ knowledge, there is no available share of supply data for these alternative technologies. However, NXP expects that its overall share of supply would […]".

See responses to question 13 of Q1 – Questionnaire to device OEMs; responses to question 16 of Q3 – Questionnaire to NFC competitors; non-confidential response by Samsung to question 16c of RFI 28, [Doc ID: 2671].

See non-confidential response by Samsung to 16.1 of Q3 – Questionnaire to NFC competitors, [Doc ID: 1095]; non-confidential response by Samsung LSI to question 15.1 of Q2 – Questionnaire to baseband competitors, [Doc ID: 1089].

(176) The combined solution can either be a single die or a stacked die solution. According to the Notifying Party, only stacked die solutions (or system-in-package; "SiP") are currently available on the market, with single die solutions are still in development, including by […] .

(177) Combined solutions comprise, on the one hand, solutions where all three components (NFC, SE, SE OS) are supplied and developed in-house by one supplier (NXP being the only such supplier today), and, on the other hand, mix-and-match solutions with components supplied and developed by two or three separate suppliers. Such mix-and-match solutions are considered by most market participants that expressed a view on the matter as viable alternatives to NXP's combined solution .

(178) Component manufacturers comprise STMicroelectronics (which manufactures NFC and SE chips), Infineon (SE chips), Oberthur (SE OS), G&D (SE OS), and Gemalto (SE OS). Apart from its integrated NFC/SE chip, NXP also supplies stand-alone NFC chips (but does currently not sell stand-alone SE chips). In addition, as mentioned in recitals (139) and (168), each of Samsung and Huawei also has in-house supply capabilities for NFC and SE chips (Samsung) or SE chips (Huawei).

(179) Suppliers of stand-alone components can either integrate their components with those from other suppliers and subsequently sell the combined solution to device OEMs, or sell their own components to third parties for integration, with the combined solution subsequently sold to device OEMs by the integrator. Such a set-up may be one where the SE supplier sells chips to the SE OS supplier that in turn sells the combined SE and SE OS to a NFC chip supplier that ultimately sells the combined NFC/SE solution to device OEMs .

(180) Device OEMs can opt for a mix-and-match solution, comprising of the components from various suppliers; alternatively they can opt for the NXP's combined solution.

(181) Device OEMs may exclusively procure the combined NFC/SE solution from NXP; others may exclusively procure a mix-and-match solution. Yet other device OEMs procure both a combined solution from NXP and a mix-and-match solution from competing suppliers. In-house developed components may also be used by device OEMs in the mix-and-match solution they adopt.

(182) For the purposes of this Decision, the Commission considers that separately from the markets for stand-alone NFC chips and stand-alone SE chips (including the SE OS) a separate product market can be identified for combined NFC/SE (including the SE OS) solutions. The competitive assessment will also be carried out on this basis.

See non-confidential response by G&D to question 16 of Q3 – Questionnaire to NFC competitors, [Doc ID: 742].

See Form CO, paragraph 870.

See non-confidential version of responses to question 29 of Q9 – Questionnaire to device OEMs; non-confidential replies to question 19 of Q11- Questionnaire to NFC competitors.

See for example non-confidential response by Gemalto to question 1.8 of RFI 25, [Doc ID: 2152]; non-confidential response by Gemalto to question 16 of Q3 – Questionnaire to NFC competitors, [Doc ID: 678].

5.2.3.3. Mobile Audio – speech enhancement software and smart amplifiers

(183) In the mobile audio segment, the Parties overlap in relation to speech enhancement software and smart amplifier chips. Those two products provide support to audio and voice functionalities within mainly smartphones and other mobile devices.

(184) Speech enhancement solutions are software-based solutions that provide calling enhancements for mobile devices. Such enhancements include noise suppression and echo cancellation. The software can be installed on the application processor, digital signal processor, Systems on a Chip ("SoC"), audio codecs, and baseband modems in mobile devices (and other products such as smart speakers, infotainment products and smart cameras).

(185) Smart amplifier chips are hardware-based discrete components that amplify a signal to allow for a better quality "speakerphone" audio transmission. Smart amplifier chips are placed between the application processor, the audio codec and the speaker in a mobile device.

(186) Qualcomm supplies its speech enhancement software product 'Fluence' […]. Qualcomm's software is available for integration on the company's Snapdragon processors and some recent MDM baseband chipsets. Through its Hexagon Access Program ("HAP"), Qualcomm facilitates the use of rival suppliers' speech enhancement software solutions with its baseband chipsets.

(187) Since 2015, Qualcomm also supplies smart amplifier chips ('Aqstic') as part of its integrated baseband chipset business. Aqstic is not supplied on a stand-alone basis, and is only available with the latest generation Snapdragon processors (namely, Snapdragon 820 and later generations). Qualcomm explains that many of its high-end Snapdragon platforms use a smart amplifier chip from rival suppliers (namely Maxim or Texas Instruments).

(188) NXP supplies a stand-alone speech enhancement software solution named 'LifeVibes', which is marketed to device OEMs, and is compatible with Qualcomm baseband chips.

(189) As for smart amplifier chips, NXP offers two different solutions for mobile devices, namely DSP-based turnkey solutions and DSP-less architecture smart amplifiers, which are sold on a stand-alone basis and can be integrated with other components of mobile devices, namely the speakers, audio codecs and baseband chips from third party suppliers, including Qualcomm.

A) Speech enhancement software

1. Notifying Party’s views

(190) With regard to speech enhancement software, the Notifying Party refers to the Commission's previous decisions, in which it considered that software markets could be segmented on the basis of (i) the different functionalities of the software and the sector concerned, and (ii) the end uses offered by the particular software. With particular regard to functionality, the Commission considered a distinction between the following types of software: (i) infrastructure software (that is to say, servers and databases), (ii) middleware (that is to say, integration platforms), (iii) application software and office software, and (iv) operating/browser software. The Commission

See Commission decision of 15 December 2014 in case M.7458 – IBM / INF Business of Deutsche Lufthansa; Commission decision of 20 June 2011 in case M.6237 – Computer Sciences Corporation / iSOFT Group.

also made a further distinction between consumer and business software within the application software segment, and between high-end, mid-range and low-end software .

(191) The Notifying Party considers that, based on the Commission's previous decisions, speech enhancement software should be considered to be a type of application software for consumers as it provides a particular range of functionalities and a further distinction between high-, mid-, and low-end software could be made based on the technology and functionalities offered.

(192) The Notifying Party considers, however, that the precise product market definition in relation to speech enhancement software can be left open as the Transaction would not raise any competition concerns.

2. The results of the market investigation and the Commission's assessment

(193) With regard to speech enhancement software, the phase I market investigation supports the notion that speech enhancement software solutions are a distinct product and perform a distinct function from other components within a mobile device .

(194) The Commission therefore considers that speech enhancement software constitutes a separate product market, distinct from other components in a mobile device, including smart amplifier chips, and from other software products performing different functions. For the purposes of this Decision, a narrower product market consisting of speech enhancement software for mobile devices incorporating a Qualcomm baseband chipset can also be considered, but whether such further segmentation is warranted can be left open as the Transaction would not raise any concerns as to its compatibility with the internal market in either case.

B) Smart amplifiers

1. Notifying Party's views

(195) With regard to smart amplifier chips, the Notifying Party refers to the Commission's decisional practice to segment semiconductor devices by type of semiconductors , and argues that smart amplifier chips can be qualified as an analog or logic IC, which can further be categorised as an Application Specific Standard Product ("ASSP") for mobile/communication devices. The Notifying Party explains further that smart amplifiers may also include a digital signal processor.

(196) The Notifying Party considers, however, that the precise product market definition in relation to smart amplifiers can be left open as the Transaction would not raise any competition concerns.

See Commission decision of 20 June 2011 in case M.6237 Computer Sciences Corporation/iSOFT Group; Commission decision of 26 January 2011 in case M.5984 Intel/McAfee; Commission decision of 20 July 2010 in case M.5904 SAP/Sybase.

See responses to question 16 of Q1- Questionnaire to device OEMs; responses to question 10 of Q4 – Questionnaire to mobile audio competitors.

See Commission decision of 17 September 2015 in Case M.7585 – NXP Semiconductors/ Freescale Semiconductor.

2. The results of the market investigation and the Commission's assessment

(197) The phase I market investigation indicated that smart amplifier chips constitute a distinct product performing a distinct function from other components within a mobile device .

(198) The Commission therefore considers that smart amplifier chips constitute a separate product market, distinct from other components in a mobile device, including speech enhancement software. For the purposes of this Decision, a narrower product market consisting of smart amplifier chips for mobile devices incorporating a Qualcomm baseband chipset could also be identified, but whether such further segmentation is warranted can be left open as the Transaction would not raise any concerns as to its compatibility with the internal market in either case.

5.2.4. Transit service technologies

(199) In the transit service technology segment, NXP is active with its proprietary contactless security technology platform MIFARE. This technology is compatible with NFC.

(200) Apart from transit ticketing and fare collection, MIFARE can be used in other (similar) applications such as in the hospitality sector (namely for hotel room doors), access cards in companies and governments, campus cards in higher education institutions. MIFARE products are used as credentials in more than 40 end-user applications.

(201) The Notifying Party explains that there are different version of MIFARE which are designed for different applications and security needs, the most important ones being MIFARE Classic, MIFARE Plus, and MIFARE DESFire. All three versions can be used in connection with NFC in mobile devices. Other NFC compatible platforms are FeliCa and CIPURSE. On mobile devices, MIFARE can be installed on the embedded SE, but also on embedded Universal Integrated Circuit Cards ("UICCs") and via replaceable SIM cards .

(202) The Notifying Party explains that MIFARE operates on the ISO/IEC 14443 RF standard using 13.56 MHz frequency. This same frequency is used by NFC to communicate. Because of the use of this same RF standard, it is explained that

See responses to question 15 of Q1- Questionnaire to device OEMs; responses to question 9 of Q4 – Questionnaire to mobile audio competitors.

The Notifying Party presents the three main versions of MIFARE as follows: MIFARE Classic is the first MIFARE version, introduced more than two decades ago. This first version of MIFARE has been compromised (its encryption can be broken and hacked) and is consequently not recommended for security-type applications, but is still used in legacy systems that rely on a MIFARE Classic infrastructure, like public transport and non-security type access applications (for example in hospitality), and occasionally in new systems with low security requirements; MIFARE Plus is a subsequent, more secure version of MIFARE that can run in two security modes, namely a standard security mode (compatible with MIFARE Classic), and a higher security mode; and MIFARE DESFire having the additional benefit of being able to perform more than one application on the same chip. It is designed for multi-application functionality in transport, micropayment, access management, and identification solutions. Another version is MIFARE Ultralight, designed for limited use, but high volume applications and described as "ideal for low-cost, high-volume applications such as public transport, loyalty cards and event ticketing, serving as the perfect contactless replacement for magnetic stripe or barcode, addressing the trend of switching entire systems to purely contactless solutions" and counting limited use tickets in public transport, event ticketing and loyalty and closed loop payment schemes as key applications; see Form CO, paragraph 770 and https://www.mifare net/en/products/chip-card-ics/mifare-ultralight/, [Doc ID 3282].

(203) For the purposes of public transport ticketing and fare collection, the Notifying Party explains that the infrastructure of a specific public transport system using MIFARE-enabled smartcards does not need to be adapted to be able to read a MIFARE-enabled mobile device as long as the same version of MIFARE is used. MIFARE, however, allows for a degree of customisation by the public transport authority or the system integrator, and this customisation must be mirrored in the MIFARE version on mobile devices (usually on the app developed by the public transport authority) to allow it to work. It is technically possible to also use a MIFARE-enabled mobile device for various different public transport systems, each transport system providing its own app for downloading on the mobile device. Competing schemes to the MIFARE, such as CIPURSE and FeliCa can also be installed on the same NFC-enabled mobile device.

(204) NXP licenses its MIFARE technology to competing suppliers of NFC/SE products and includes it on […] of the NFC/SE products it ships to customers .

(205) Qualcomm is not active in this market segment.

5.2.4.1. Notifying Party's views

(206) The Notifying Party does not provide an assessment of MIFARE for the purposes of product market definition. It does, however, indicate that apart from MIFARE, other transport ticketing technologies are also available for similar use cases, especially mobile ticketing and fare collection, in NFC compatible mobile devices. According to the Notifying Party, such other technologies comprise the CIPURSE open securing standard, the FeliCa technology by Sony, and the Calypso contactless ticketing system developed by a number of European transport operators and authorities. In addition, the Notifying Party indicates QR Codes, or the "open loop" EMV based mobile wallets (such as Apple Pay, Android Pay, Samsung Pay) as alternatives to MIFARE, giving the London Underground as an example of the deployment of EMV in parallel to MIFARE .

(207) Furthermore, the Notifying Party explains that to date the utilisation of MIFARE on mobile devices has been low and that while it is expected that MIFARE in NFC-enabled mobile devices will eventually be used in applications for which MIFARE-enabled smartcards are currently used, such as transit and access control, and despite a large number of NFC/SE products currently shipped containing MIFARE, that capability is currently not being used by the main mobile device OEMs. The Notifying Party explains that while embedded SE chips may be MIFARE enabled, this does not mean that MIFARE is implemented in the mobile device. Currently no mobile devices containing MIFARE enabled embedded SEs can be used for mobile transit .

See Form CO, paragraph 777.

The Notifying Party explains that in 2016, NXP shipped […] NFC/SE units which were MIFARE enabled; [40-50]% of NXP's NFC/SE shipments included a MIFARE implementation, including […] NFC/SE shipments intended for Android devices; see Form CO, paragraph 779.

5.2.4.2. The results of the market investigation and the Commission's assessment

(208) The market investigation indicates the existence of alternative technologies to MIFARE for transit ticketing (including CIPURSE, Calypso). However, many amongst the respondents consider that MIFARE has specific advantages over other technologies employed for public transport. Respondents mention reliability, security, execution speed, and implementation and maintenance costs as key characteristics of MIFARE. Key market players (namely Samsung, Gemalto, G&D) make reference to the wide deployment, installed base and high switching costs for transport authorities as elements distinguishing MIFARE from other competing technologies .

(209) Thales, a solution provider for fare collection in public transportation, summarises the various technologies as follows: "Mifare and CiPurse are both designed for a compromise between security strength, cost effectiveness and speed performance, which are key requirements for mass transit markets. QR code is very cheap but has inherent security weakness and can sometimes be very slow to read. EMV is very secure, but pretty slow for mass transit. O[n] the top of that using EMV cards requires complying with banking security standards, which makes the overall cost of implementation high and the return on investment debatable for many transport operators" .

(210) The responses to the market investigation also indicate that while mobile ticketing and fare collection via smartphone may still be limited today, transport companies are expecting such a move in the next few years. Mobile devices are increasingly MIFARE enabled and the attach rate in respect of MIFARE is expected to increase, and more so than competing technologies such as Calypso or CIPURSE, in the next two to three years .

(211) When asked about expectations as to the development of usage and instalment of FeliCa, Calypso and other technologies on smartphones in the next two-to-three years, as compared to MIFARE technology, the responses by market participants suggest an increased importance for MIFARE. Oberthur notes that "MIFARE is de

See response by the Notifying Party of 28 June 2017 to the Article 6(1)(c) decision, paragraph 106, [Doc ID 1331].

According to ABI Research presentation "Transportation Ticketing Technologies Market Update" of 2 January 2015 CIPURSE "[…]", [Doc ID: 1334-6854].

See responses to Q12 – Questionnaire to Public transit authorities; responses to Q14 – Questionnaire to Reader infrastructure integrators.

See responses to question 51 of Q3 – Questionnaire to NFC competitors.

See non-confidential response by Thales to question 6.1 of Q14 – Questionnaire to reader infrastructure integrators, [Doc ID: 2037].

See non-confidential responses to question 9 of Q12 – Questionnaire to Public transport authorities.

See non-confidential responses to question 49 of Q1 – Questionnaire to device OEMs; non-confidential responses to question 42 of Q2 – Questionnaire to baseband competitors; non-confidential responses to questions 48-50 of Q3 – Questionnaire to NFC competitors.

See non-confidential response by Oberthur to question 48.1 of Q3 – Questionnaire to NFC competitors, [Doc ID: 761].

See non-confidential response by G&D to question 48.1 of Q3 – Questionnaire to NFC competitors, [Doc ID: 742].

See non-confidential response of Gemalto to question 48.1 of Q3 – Questionnaire to NFC competitors, [Doc ID: 678].

See non-confidential response by Gemalto to question 46.1 of Q3 – Questionnaire to NFC competitors, [Doc ID: 678].

facto a dominant technology in transport, G&D refers to the "[c]lose to 80% market share in public transport", while Gemalto explains that MIFARE "is and will remain the dominant technology while FELICA and Calypso will be fringe technologies used on very few public transport systems" .

(212) Gemalto further explains the importance of MIFARE for smartphones as follows: "[t]he fact that more and more consumers will want to use smartphones for contactless payment including for public transportation means that smartphone OEMs will want to ensure that their devices are NFC-enabled and will support MIFARE by including an MIFARE-enabled SE in their devices" .

(213) Various technologies therefore exist in the transit service technology space, including MIFARE, Calypso, and CIPURSE. Those technologies may, however, not be fully substitutable from a demand-side perspective, neither at the level of device OEMs that incorporate such technologies on the mobile devices to allow for mobile ticketing, nor at the level of local transport authorities to the extent that such authorities have already opted for one technology for contactless ticketing in their network. Conversely, when a transit service system is set up, a transport authority would have various options to choose from, including MIFARE, Calypso, CIPURSE. Some market participants have argued, nonetheless, that MIFARE and other technologies such as FeliCa and Calypso may be complementary rather than substitutes from a device OEMs perspective. Gemalto explains that device OEMs do "want their smartphones to support contactless payment systems such as FELICA and Calypso that are used on some public transport systems, but this is always in addition to MIFARE", whereas G&D notes that "Felica, Calypso and other [t]echnologies are to be considered as complementary […] not resulting in the [s]ubstitution […] of the requirement for Mifare" .

(214) The phase II market investigation has also shown that it is not common for transit authorities to change transit service technology given the significant switching costs .

See responses to questions 7-8 of Q12 – Questionnaire to Public transit authorities.

(215) […] .

(216) Gemalto […] notes that "[t]he rollout of contactless payment to a city's transport system requires significant technology-specific investments. For travel[l]ers to be able to pay via their smartphones, readers need to be installed that are equipped with chips on which MIFARE is installed. In addition to the cost of readers; approval procedures, testing and other installation costs further contribute to the cost of setting up a contactless payment technology. As such, switching to an alternative technology would be very costly. Alternative technologies for contactless payment on public transportation exist and are used in some cites (e.g. Calypso, which is used in Brussels, Paris, Lisbon, Milan and certain other cities in Europe; or FeliCa which is used in Japan and Hong Kong). However the large technology-specific investments required to implement a contactless payment system (amounting to several tens of million Euros per city) mean that cities that have already opted for MIFARE are unlikely to switch to alternative contactless technologies any time soon." .

See non-confidential response by Gemalto to question 49.1 of Q3 – Questionnaire to NFC competitors, [Doc ID: 678].

See non-confidential response by Thales to question 13.1 of Q14 – Questionnaire to Reader infrastructure integrators, [Doc ID 2037].

See responses to question 47 of Q1 – Questionnaire to device OEMs; responses to question 41 of Q2 - Questionnaire to baseband competitors; responses to question 46 of Q3 – Questionnaire to NFC competitors.

See responses to questions 50-51 of Q1 – Questionnaire to device OEMs; responses to question 49 of Q3 – Questionnaire to NFC competitors; responses to question 50 of Q3 - Questionnaire to NFC competitors.

(220) Patents that are essential to a standard are those that cover technology to which a standard makes reference and that implementers of the standard cannot avoid using in standard-compliant products. Those patents are known as standard-essential patents (SEPs). SEPs are different from patents that are not essential to a standard ("non-SEPs"). This is because it is generally technically possible for an implementer to design around a non-SEP to comply with a standard. By contrast, an implementer typically has to use the technology protected by a SEP when manufacturing a standard-compliant product.

(221) In previous decisions , the Commission concluded that each SEP represents a separate relevant technology market, as it is necessary to comply with a standard and thus cannot be designed around. By definition, there is no alternative or substitute for a SEP. As explained in Google / Motorola Mobility, "[a] company wishing to produce goods complying with a certain standard cannot do so without either a licence to the technology incorporated in that standard or by infringing the patents covering that technology. Prior to the adoption of a standard, multiple technologies may have competed. However, once a standard has been adopted and widely implemented by the industry and in the absence of competing standards, firms that use these technologies may be severely limited in their ability to use another technology. The very purpose of choosing a standard is that the industry coordinates on a specific technological solution at the expense of alternative technologies. Inter-technology competition that existed before is therefore impeded and any alternative technologies or technical solutions that may have had the same functionalities as the one chosen as the standard technology may have a significantly reduced value. In other words, once the standard is set, and in the absence of a competing standard, technology competition is largely eliminated. […] The specificity of SEPs is that they have to be implemented in order to comply with a standard and thus cannot be designed around, i.e. there is by definition no alternative or substitute for each such patent. Therefore, each SEP constitutes a separate relevant technology market on its own." .

See Commission decision of 13 February 2012 in Case M.6381 – Google / Motorola Mobility;

See Commission decision of 4 December 2013 in Case M.7047 – Microsoft / Nokia; Commission decision of 29 April 2014 in Case AT.39939 – Samsung. Enforcement of UMTS standard essential patents; Commission decision of 29 April 2014 in Case AT.39985 – Motorola – enforcement of GPRS standard essential patents.

(222) With regard to non-SEPs, the Commission's decision in Google / Motorola Mobility noted as follows: "As regards non-SEPs, the commercial importance of these patents varies. Such patents are not part of a formal technical standard, the nature of many such patents may be incremental, and it is often easier to design around a patent falling in this category. Non-SEPs may relate to features used to differentiate competitors' products on the market, thus creating dimensions on which firms aggressively compete. This being said, non-SEPs could also potentially be the basis for foreclosure of rivals and possible abusive conduct. For example, in exceptional circumstances, notably where a technology has become an indispensable input for competitors, a refusal to grant access to that technology may be abusive." .

5.2.5.1. Notifying Party's views

(223) The Notifying Party refers to point 22 of the Technology Transfer Guidelines for the purposes of product market definition in the context of patents and licensing, which reads: "The relevant technology markets consist of the licensed technology rights and its substitutes, that is to say, other technologies which are regarded by the licensees as interchangeable with or substitutable for the licensed technology rights, by reason of the technologies' characteristics, their royalties and their intended use. Starting from the technology which is marketed by the licensor, it is necessary to identify those other technologies to which licensees could switch in response to a small but permanent increase in relative prices, that is to say, to the royalties. An alternative approach is to look at the market for products incorporating the licensed technology rights […]." .

(224) The Notifying Party considers that on a general level and for market definition purposes, licensed patents should be considered as a family rather than as individual patents under national law. This is especially the case in sectors where both licensing and competition in downstream product markets takes place at a supra-national level, which is the case in the semiconductor and mobile device industries. While recognising that a single invention can be patented under various different national systems, it nonetheless remains a single invention. This would not mean, however, that each patent family would form its own relevant product market .

(225) The Notifying Party refers to the Commission's decisions in relation to market definition of SEPs, namely that each SEP constitutes a separate product market, but refrains from endorsing such an approach. Instead, the Notifying Party puts forward observations on the implications that such an approach on market definition would necessarily have. The Notifying Party contends that if indeed each SEP constitutes a separate product market, the SEP holder will be a monopolist on those markets, and consequently a combination of SEPs cannot give rise to any horizontal overlaps: there is no increment of market share when the SEP holder is already a monopolist with 100% market share on the market. With reference to the Commission's previous decisions, the Notifying Party also pinpoints the alleged dominant position that each holder of a SEP would have in respect of the SEPs it holds, and to the significant market power that the mere fact of holding such SEPs may give rise to.

(226) The Notifying Party notes, however, that despite such decisions, NXP does not have market power in respect of NFC or NFC SEPs (as evidenced, according to the Notifying Party, by the responses to the phase I market investigation). According to the Notifying Party there is therefore a "clear tension" between the Commission's approach to market definition and the assessment on the market power of NXP in respect of its NFC IP, which leads to conclude that the Commission's approach to market definition in respect of SEPs may, in fact, be incorrect .

(227) If each SEP constitutes a separate market, this would exclude the possibility of there being a relevant market for all NFC patents or for an NFC patent portfolio to the extent such a portfolio would include SEPs (which is the case for NXP and Qualcomm).

(228) The Notifying Party suggests that the relevant product market could be wider, and include not only NFC patents but also patents reading on other technologies which are viable substitutes to NFC. This could be the case considering the strong inter-technological competition and the existence of viable substitutes to NFC technology and, by extension, viable substitutes to NFC patents including NFC SEPs.

(229) With regard to non-SEPs, the Notifying Party considers that while all non-essential NFC patents could hypothetically form a single product market, this would be a "crude and meaningless classification" which would not take account of substitutability or the lack thereof between patents and groups of patents. Alternatively, non-essential NFC patents could be categorised by "their broad technological function", namely "Air interface (Inter-device)", "Device aspects (Intra-device)", "Use cases/services", and "Security", but such categorisation would again not be meaningful as many patents within each category would be complements rather than substitutes. The Notifying Party also considers the notion of competition between NFC portfolios as unconvincing, as portfolios of licensed IP including NFC patents, could be seen as complements rather than substitutes.

See response by the Notifying Party to question 6 of RFI 34 [DOC ID 02333].

See response by the Notifying Party of 28 June 2017 to the Article 6(1)(c) decision, [Doc ID 1331]; response by the Notifying Party to question 6 of RFI 34 [DOC ID 02333].

See response by the Notifying Party to questions 1.c.i and 6 of RFI 34.

(230) The Notifying Party considers that the precise approach to product market definition in respect of NFC patents can be left open as no concerns arise under any plausible market definition .

5.2.5.2. The results of the market investigation and the Commission's assessment

(231) The results of the phase I and phase II market investigation indicate that the Commission's conclusions in its previous decisions are also valid for NXP's and Qualcomm's SEPs for NFC and cellular technologies respectively . One respondent notes that […] .

(232) With regard to other IP that does not qualify as SEPs, the Commission notes that both Qualcomm and NXP possess IP that is relevant for the purpose of NFC technology. Given that NFC technology can be viewed as a separate product market, the Commission considers that non-SEP IP related to NFC technology may be viewed as a distinct product market from non-SEP IP related to other technologies (as well as from SEP IP related to NFC). The Commission notes, however, that contrary to SEPs, a non-SEP for a given functionality/product may in theory be substitutable with another non-SEP which allows making that same functionality/product or even with another non-SEP which allows making a substitutable functionality/product.

(233) The phase I market investigation generated mixed responses to the question as to whether non-SEP IP related to NFC technology could be designed around or alternative technologies could be used to achieve the same functionality. Intel explains that "[i]t is impossible to design around all NFC non-SEPs. One reason for this is that certain core legacy technologies that are part of NFC are covered by non-SEPs. These patents cover basic aspects of NFC technology that are so fundamental that it would be impossible to pass an NFC compliance test without infringing on at least some of them. For example, the manner in which an NFC chipset works with an SE chipset is a core aspect of NFC technology, and is patented but not necessarily standardized. […] In addition, it may be impossible to produce a competitive chip from the standpoint of power, performance, or form factor without infringing non-SEP differentiating patents.". Conversely, Microsoft explains that non-SEPs in the software space can be worked around, and would "only achieve market power by virtue of competition on the merits with other solutions" .

(234) In any event, the market investigation did not provide any evidence that NXP's NFC non-SEPs are commercially essential.

(235) The Commission also notes that NXP licenses its NFC IP to third parties not as single patents but as a group of patents. It has […]. In addition it has […] . NFC has also […] .

See response by the Notifying Party to question 6 of RFI 34 [DOC ID 02333].

See responses to question 14 of Q1 – Questionnaire to device OEMs; responses to question 16 of Q2 – Questionnaire to baseband competitors; responses to question 17 of Q3 – Questionnaire to NFC competitors.

See response by […] to question 14 of Q1 – Questionnaire to device OEMs, [Doc ID: 1009].

See non-confidential response by Intel to question 14.2 of Q11- Questionnaire to NFC competitors, [Doc ID: 2303].

See non-confidential response by Microsoft to question 14.1 of Q1 – Questionnaire to device OEMs, [Doc ID: 776].

See responses to question 16 of Q11 – Questionnaire to NFC Competitors.

(236) The Commission considers that, in line with its previous decisions, each SEP related to cellular and NFC technology should be considered as a separate market. The Commission also considers that non-SEP IP related to NFC technology may be viewed as a distinct product market from non-SEP IP related to other technologies. The exact product market definition for non-SEP IP relevant for the purpose of NFC technology is, however, left open.

5.3. Geographic market definition

5.3.1. All relevant semiconductor products (automotive semiconductors, IoT semiconductors, baseband chipsets, NFC, SE, and combined NFC/SE chips, smart amplifiers)

5.3.1.1. Notifying Party's views

(237) The Notifying Party recalls that the Commission has considered the geographic scope of semiconductor markets in its previous decisions and concluded that the markets may be at least EEA-wide, if not worldwide, although the precise geographic scope of those markets was left open .

(238) The Notifying Party submits that the geographic market definition in the Commission's previous cases, notably in NXP/Freescale, should also be retained in this Decision and that the relevant geographic market for all semiconductors should be considered as worldwide in scope. It is further argued that as market conditions in the semiconductor industry have not changed materially, there is no justification to define narrower markets than the worldwide market. In fact, suppliers of semiconductors conduct business on a worldwide basis, with recourse to manufacturing plants located around the world; suppliers compete on a worldwide level; customers have cross-border sourcing strategies and also compete on a worldwide level; no regulatory barriers to trade of semiconductors on a worldwide level exist, there are no quotas, tariffs, or technical specifications that could hinder cross-border trade; and transportation costs remain low, with minimal impact on pricing.

(239) The Notifying Party also notes that it is not aware of any EEA-specific variations in relevant product segments that would mean that shares of supply at the EEA level would differ materially from the Parties' and their competitors' worldwide shares of supply. Consequently, worldwide shares of supply could also be broadly representative of the market position of the Parties and their competitors at the EEA level .

See copies of license agreements with […] provided as Annex 4.15.5 and Annex 4.15.6 to the Form CO.

See copies of license agreements with […] provided as Annex 4.15.3, Annex 4.15.1 and Annex 4.15.4 to the Form CO.

See copy of license agreement with […] provided as Annex 4.15.2 to the Form CO.

5.3.1.2. The results of the market investigation and the Commission’s assessment

(240) In NXP/Freescale , based on the market investigation, the Commission found that the geographic scope of the semiconductor markets was likely to be worldwide in scope, as competition between suppliers is worldwide, transport costs are very low, and price differences among regions are small. The exact market definition was, however, ultimately left open.

(241) The market investigation in this Decision provides strong indications that the various possible semiconductor markets are likely to be worldwide in scope. This is also in line with the Commission's previous decisions, in particular NXP/Freescale .

(242) For the purposes of this Decision, the Commission therefore concludes that the geographic scope of the semiconductor product markets relevant in this case, including semiconductors for automotive applications, semiconductors for IoT applications, and semiconductors for mobile devices (baseband chipsets, NFC, SE, and combined NFC/SE chips, and smart amplifiers) is likely worldwide.

5.3.2. Speech enhancement software

5.3.2.1. Notifying Party's views

(243) In its previous decisions, the Commission considered the market for software products and took the view that the geographic scope of the market was at least EEA-wide, but ultimately left the exact geographic market definition open.

(244) The Notifying Party agrees with that assessment and considers that the market for software (including speech enhancement software) is worldwide in scope based on the same justifications as for semiconductor devices, including the fact that software may be uploaded/downloaded anywhere in the world and transport costs are negligible or non-existent.

(245) The Notifying Party also notes that it is not aware of any EEA-specific variations that would mean that shares of supply at EEA level would differ materially from the Parties' and their competitors' worldwide shares of supply for speech enhancement software. Consequently, worldwide shares of supply could also be broadly representative of the market position of the Parties and their competitors at the EEA level .

5.3.2.2. The results of the market investigation and the Commission's assessment

(246) With regard to software products, the market investigation indicates that device OEMs source speech enhancement solutions from both vendors located in the EEA

See Commission decision of 17 September 2015 in case M.7585 – NXP Semiconductors / Freescale Semiconductor.

See responses to question 18 of Q1 – Questionnaire to device OEMs; responses to question 17 of Q2 – Questionnaire to baseband competitors; responses to question 18 of Q3 – Questionnaire to NFC competitors; responses to question 12 of Q4 – Questionnaire to mobile audio competitors; responses to questions 22-23 of Q5 – automotive competitors; responses to questions 23-24 of Q6 – Questionnaire to automotive customers; responses to question 12 of Q7 – Questionnaire to IoT competitors; responses to question 12 of Q8 – Questionnaire to IoT customers.

See Commission decision of 17 September 2015 in case M.7585 – NXP Semiconductors / Freescale Semiconductor.

See Commission decision of 15 December 2014 in case M.7458 – IBM / INF Business of Deutsche Lufthansa; Commission decision of 20 July 2010 in case M.5904 SAP/Sybase.

See response by the Notifying Party to questions 1-2 of RFI 54.

and in the rest of the world. Similarly, suppliers of speech enhancement software supply customers on a worldwide basis .

(247) The exact geographic market definition can, however, be left open as the Transaction does not raise competition concerns with respect to speech enhancement software irrespective of the exact geographic market definition.

5.3.3. Transit service technologies

5.3.3.1. Notifying Party's views

(248) The Notifying Party does not provide an assessment of MIFARE or transit service technologies for the purposes of geographic market definition. It notes, however, that competition takes place on a worldwide basis .

5.3.3.2. The results of the market investigation and the Commission's assessment

(249) The various leading transit service technologies such as MIFARE, FeliCa, and Calypso are offered on a worldwide basis, although deployed to various degrees in different regions. MIFARE is the leading worldwide transit service technology, deployed in 750 cities worldwide. FeliCa is primarily used in Japan, but has also been deployed for transit systems elsewhere, including in Hong Kong, India, Indonesia, Singapore. Calypso, initially developed by a number of European transit operators, is deployed across 125 cities in 25 countries, with primary focus on certain European countries (including France, Belgium, Italy) but also in Israel, Canada, Mexico, and several North African countries .

(250) Customers of transit system technologies comprise transit authorities worldwide, as well as device OEMs to the extent that such technologies are included on the mobile devices for mobile ticketing purposes. Similarly to other components, device OEMs procure transit system technologies, including in particular NFC/SE solutions that are MIFARE enabled, on a worldwide basis, while MIFARE enabled NFC/SE solution providers offer their MIFARE enabled products to customers worldwide .

(251) The Commission concludes, therefore, that the market for transit service technologies is likely worldwide.

5.3.4. Intellectual Property (SEPs and non-SEPs)

5.3.4.1. Notifying Party's views

(252) For the purposes of geographic market definition, the Notifying Party refers to Article 1(1)(l) of the Technology Transfer Block Exemption Regulation, and point 24 of the Technology Transfer Guidelines, which reads: "The ‘relevant geographic market’ is defined in Article 1(1)(l) of the TTBER and comprises the area in which

the undertakings concerned are involved in the supply of and demand for products or the licensing of technology, in which the conditions of competition are sufficiently homogeneous and which can be distinguished from neighbouring areas because the conditions of competition are appreciably different in those areas The geographic market of the relevant technology market(s) can differ from the geographic market of the relevant product market(s)" .

See response by the Notifying Party to questions 1-2 of RFI 54.

(253) The Notifying Party also refers to the Commission's decisional practice concerning the geographic scope of the market for the licensing of (cellular) SEPs, which the Commission has considered to be at least EEA wide.

(254) With regard to NFC chips, the Notifying Party argues that this technology is implemented throughout the EEA, but is also implemented to some extent in other jurisdictions around the world. Considering that it is common practice in the industry to grant worldwide licenses for patent portfolios, the Notifying Party concludes that the geographic scope of the relevant market for non-SEP IP related to NFC to be worldwide in scope.

(255) Nevertheless, the Notifying Party considers that the exact definition can be left open with respect to NFC patents as the Transaction would not raise any concerns under any plausible market definition .

5.3.4.2. The results of the market investigation and the Commission's assessment

(256) In its previous decisions, the Commission considered the market for the licensing of SEPs as being at least EEA-wide . For the purposes of this decision, the Commission considers this conclusion to remain valid.

(257) Similarly to SEPs, also the market for the licensing of non-SEP IP relevant for the purpose of NFC technology is likely to be at least EEA-wide. The Commission notes, in this context, that NXP has entered into worldwide NFC IP licensing agreements with competitors ([…]) and customers ([…]) active in all parts of the world. NXP's MIFARE licensing agreements have similarly also a worldwide scope and have been entered into with market players active in various parts of the world.

(258) For the purposes of this decision, the Commission considers that the geographic market definition for non-SEP IP relevant for the purpose of NFC technology can be left open.

6. COMPETITIVE ASSESSEMENT – HORIZONTAL NON-COORDINATED EFFECTS

6.1. Analytical framework

(259) The Horizontal Merger Guidelines describe horizontal non-coordinated effects as follows:

“A merger may significantly impede effective competition in a market by removing important competitive constraints on one or more sellers who consequently have increased market power. The most direct effect of the merger will be the loss of competition between the merging firms. For example, if prior to the merger one of the merging firms had raised its price, it would have lost some sales to the other merging firm. The merger removes this particular constraint. Non-merging firms in the same market can also benefit from the reduction of competitive pressure that results from the merger, since the merging firms’ price increase may switch some demand to the rival firms, which, in turn, may find it profitable to increase their prices. The reduction in these competitive constraints could lead to significant price increases in the relevant market.” .

(260) Therefore, a merger giving rise to such non-coordinated effects might significantly impede effective competition by creating or strengthening the dominant position of a single firm, one which, typically, would have an appreciably larger market share than the next competitor post-merger.

(261) However, under the substantive test set out in Article 2(2) and Article 2(3) of the Merger Regulation, also mergers that do not lead to the creation of or the strengthening of the dominant position of a single firm may create competition concerns. Indeed, the Merger Regulation recognises that in oligopolistic markets, it is all the more necessary to maintain effective competition. This is in view of the more significant consequences that mergers may have on such markets. For this reason, the Merger Regulation clarifies that "under certain circumstances, concentrations involving the elimination of important competitive constraints that the merging parties had exerted upon each other, as well as a reduction of competitive pressure on the remaining competitors, may, even in the absence of a likelihood of coordination between the members of the oligopoly, result in a significant impediment to effective competition" .

(262) The Horizontal Merger Guidelines list a number of factors which may influence whether or not significant horizontal non-coordinated effects are likely to result from a merger, such as the large market shares of the merging firms, the fact that the merging firms are close competitors, the limited possibilities for customers to switch suppliers, or the fact that the merger would eliminate an important competitive force. That list of factors applies equally regardless of whether a merger would create or strengthen a dominant position, or would otherwise significantly impede effective competition due to non-coordinated effects. Furthermore, not all of those factors need to be present to make significant non-coordinated effects likely and it is not an exhaustive list. Finally, the Horizontal Merger Guidelines describe a number of factors, which could counteract the harmful effects of the merger on competition, including the likelihood of buyer power, entry and efficiencies.

Horizontal Merger Guidelines, paragraph 24.

Merger Regulation, recital 25.

Merger Regulation, recital 25. Similar wording is also found in paragraph 25 of the Horizontal Merger Guidelines. See also Commission decision of 2 July 2014 in case M.7018 – Telefónica Deutschland/E-Plus, recital 113; Commission decision of 28 May 2014 in case M.6992 – Hutchison 3G UK/Telefónica Ireland, recital 179; Commission decision of 12 December 2012 in case M.6497 – Hutchison 3G Austria/Orange Austria, recital 88.

(263) The Commission concludes that the increment brought about by the Transaction would be minor (that is to say, approximately [0-5]%).

(268) Moreover, other competitors would remain in the market post-Transaction including established market players with high market shares, such as Renesas and Texas Instruments, and disruptive new entrants, such as Nvidia. The presence of alternative providers post-Transaction has been confirmed by the phase I market investigation. The majority of customers and competitors responding to the phase I market investigation confirm that a sufficient number of manufacturers and suppliers would remain active post-Transaction irrespective of a further segmentation of the market in entry and mid/high tiers. Continental highlights how Qualcomm and NXP have complimentary portfolios; while General Motors indicates that the number of suppliers would remain adequate since new players, such as Intel and NVidia have recently also started to offer their products in the market .

(269) The phase I market investigation also indicates that additional suppliers are expected to enter into this market over the coming years including players which are traditionally active in the supply of semiconductors for mobile handsets such as Samsung and Intel .

(270) The majority of respondents further believe that the Parties are not close competitors, irrespective of a further segmentation of the market in entry and mid/high tiers. Some customers note that within the overall Infotainment MPU category NXP is more present in the entry segment and Qualcomm in the premium segment and that today Qualcomm’s market share in the market is low. This has been confirmed by certain competitors who consider Qualcomm as a limited market player with presence only in the high/premium segment .

(271) Finally, the majority of respondents do not consider that the Transaction would have an impact on the market for the manufacture and supply of Infotainment MPUs, irrespective of any further segmentation. Most of the competitors believe that enough choice would remain available to customers and that there was little overlap between the two companies. Only one respondent notes that Qualcomm, through a more complete offering, may be able to increase its position in the market and push traditional players out of the market. Among the customers responding to the phase I market investigation, the majority consider that the Transaction would have a positive impact on the market due to the complementarity of NXP and Qualcomm portfolios which will drive the market to a higher standard .

(272) The Commission therefore concludes that the Transaction would not raise competition concerns as to its compatibility with the internal market on the market for the manufacture and supply of Infotainment MPUs for automotive applications and any potential narrower markets.

See responses to question 22 of Q6 – automotive customers, responses to question 26 of Q5 – automotive competitors.

See responses to question 29 of Q6 – automotive customers, responses to question 28 of Q5 – automotive competitors.

See responses to question 22 of Q6 – automotive customers, responses to question 25 of Q5 – automotive competitors.

See responses to question 47 of Q6 – automotive customers, responses to question 46 of Q5 – automotive competitors.

61

(283) Based on the market shares illustrated in Table 3, the Parties would be the first supplier of Infotainment connectivity with a market share of [20-30]%. However, in this market segment, NXP's market share over the last years is negligible and so is the increment that would be brought about by the Transaction (that is to say, approximately [0-5]%).

(284) Moreover, other competitors would remain in the market post-Transaction including established market players, such as Toshiba, Renesas and Panasonic Corporation with all three players having relatively stable market shares of approximately [10-20]% over the last three years. The presence of alternative providers post-Transaction has been confirmed by the phase I market investigation. The large majority of respondents confirm that other manufacturers and suppliers would remain active post-Transaction. In particular, while most of the competitors note that several suppliers would remain available; customers note that, in any case, the Transaction would not change the supplier landscape .

(285) Moreover, the majority of respondents believe that the Parties are not close competitors on the market segment for Infotainment connectivity. While customers such as Valeo/Peiker, Fiat and BMV believe that the Parties' portfolios do not overlap, Renesas indicates that the Parties have complementary technology.

(286) With regard to potential new entrants on this market segment, some respondents indicate that other suppliers may enter the market, in particular Asian suppliers such as Samsung .

(287) Finally, the majority of respondents believe that the Transaction would not have any impact on the market for Infotainment connectivity for automotive applications .

(288) The Commission therefore concludes that the Transaction would not raise competition concerns as to its compatibility with the internal market on the market for the manufacture and supply of Infotainment connectivity chips for automotive applications.

6.2.5. Automotive chips based on non-cellular V2X technology

(289) As described in the Linley report, the automotive industry has been investigating methods that allow vehicle-to-vehicle ("V2V"), vehicle-to-infrastructure ("V2I"), vehicle-to-pedestrian ("V2P"), vehicle-to-network communications also known as V2X.

(290) V2X communications are processed with recourse to a chip connected to the central ADAS processor. V2X is a technology used to connect cars to various external stimuli (other cars, bikes, road works, infrastructure) and to provide the unique capability to see past other vehicles, as well as around corners, obstacles, or turns on the road. V2X technology may become an important component for the securely connected, self-driving car of the future. Two types of cellular communications have emerged: cellular V2X and non-cellular (based on the 802.11.p technology).

(291) On 30 November 2016, the Commission adopted a European Strategy (COM (2016) 766) on Cooperative Intelligent Transport Systems (“C-ITS”), a milestone initiative towards cooperative, connected and automated mobility. Moreover, since 2014, the Commission is also consulting public and private stakeholders within the C-ITS platform, to develop a shared vision on the interoperable deployment of C-ITS in the EU. The Commission is advocating for C-ITS to be implemented through a combination of ETSI ITS-G5 and existing cellular networks, through standardisation and EU-wide deployment specifications. This view was supported by respondents to the public consultation who gave widespread support for the hybrid communication approach .

See responses to question 32 of Q6 – automotive customers, responses to question 31 of Q5 – automotive competitors.

See responses to question 34 of Q6 – automotive customers, responses to question 33 of Q5 – automotive competitors.

6.2.5.1. Notifying Party's views

(292) The Notifying Party believes that there is no risk that the Transaction would harm competition for several reasons. First, there are a number of alternative suppliers that are active in the market. The leading vendors of processors in the autonomous driving technology are Mobileye, Intel, Infineon, NVidia, Texas Instruments and Toshiba. Second, in relation to the V2X technology, the Parties note that the system is a nascent segment. It is still in its early adoption phase […].

(293) Third, the Parties further note that they are not close competitors on the market. NXP’s development efforts in V2X systems have generally been directed at non-cellular based systems while, by contrast, Qualcomm focuses most of its investment on cellular based V2X technology. Moreover, other suppliers such as Autotalks or Renesas are closer competitors to NXP that Qualcomm is, while Qualcomm's closest competitors in the V2X segments are likely to be firms with cellular wireless expertise such as Intel or Huawei. Finally, the Notifying Party considers that other suppliers would start providing chips based on V2X technology in the coming years.

(294) Moreover, in response to the Article 6(1)(c) decision, the Notifying Party argues that the Transaction would not harm competition in the DSRC V2X market segment for several reasons. First, the overlap between the Parties is limited to the radio aspect of DSRC V2X solutions, meaning that the overlap is limited to a component of DSRC V2X solutions, itself a sub-segment of V2X, which is in turn a sub-segment of ADAS. Second, DSRC V2X will play a significant role, or indeed any role, in ADAS solutions in the foreseeable future. Third, barriers to entry in the supply of DSRC V2X solutions – in particular the segment in which there is an overlap – are low due to (i) the rudimentary nature of the technology involved; (ii) the numerous semiconductor players that supply 802.11-standard compliant products and have the necessary expertise to enter quickly the segment; and (iii) the standardisation context, meaning that the relevant technology is publicly available. Finally, the Transaction will not harm innovation competition.

6.2.5.2. The results of the market investigation and the Commission's assessment

(295) In the Article 6(1)(c) decision, while the Commission acknowledged that the V2X non-cellular is a nascent market segment, it also raised serious doubts as to the compatibility of the Transaction with the internal market on the grounds that the elimination of the competitive pressure between the Parties would have weakened competition in the market for the development of V2X chips based on the non-cellular technology. Moreover, the Commission also raised the question on whether the Transaction may harm the future development of V2X chips based on non-cellular technology.

See responses to question 34 of Q6 – automotive customers, responses to question 33 of Q5 – automotive competitors.

(296) Qualcomm is investing in ADAS technologies and it plans to achieve significant revenues in this market over the coming years. In particular, with regard to V2X technologies, many respondents view Qualcomm as an important developer of both cellular and non-cellular V2X chips. With regard to NXP, the market shares show that it is an important player in the overall ADAS market. NXP is also significantly investing in ADAS technologies in the next years and plans to achieve significant revenues over the next years .

A) Non-cellular V2X technologies would be important for autonomous driving and advanced safety systems

(297) With regard to the importance of V2X technology for semiconductors for automotive application, the majority of respondents to the phase I market investigation submitted that, if V2X technology is relatively important today, it will become very important in the next five years. One respondent notes that certain big OEMs announced high volume market introduction of V2X in Europe in 2019. Others consider that, in the coming years, it is expected that transportation infrastructure will have the required connectivity and when the infrastructure will be ready, V2X will be the key feature to support safer driving . Most of the respondents to the phase II market investigation confirm that V2X technology is important today and it would become very important in the next three years. The phase II market investigation provided mixed results on whether it would not be possible to have full autonomous driving system without adopting V2X technologies. While most of the customers consider V2X technology to be essential, some of the competitors note that similar safety functionalities may be achieved through the use of sensors and artificial intelligence (“AI”) technology .

(298) The growing relevance of V2X is also confirmed by documents provided by the Parties. NXP foresees V2X penetration to increase in […].

(299) Abi Research expects that by 2025, [70-80]% of the cars in Western Europe will be equipped with V2X. NXP also foresees a strong increase in the penetration rate of cars equipped with V2X technology in the EU. Based on internal estimates, by 2029, all […] cars will include V2X chips .

See Slide 66 - Annex 1.7, Form CO. In particular, Qualcomm plans to achieve revenue over […] and of approximately […] [DOC ID 327-25].

See responses to question 41 of Q6 – automotive customers, responses to question 40 of Q5 – automotive competitors.

See slides 21, 67 and 76, […] of 25 February 2016.

See slide 99, […] of 21-23 March 2016. Expected revenues to reach […].

See responses to question 40 of Q6 – automotive customers, responses to question 39 of Q5 – automotive competitors.

See responses to question 18 of Q16 – automotive customers and responses to question 13 of Q15 – automotive competitors.

See responses to question 17 of Q16 – automotive customers and responses to question 14 of Q15 – automotive competitors.

66

Figure 1: EU OEM V2X feature penetration

[…]

Source: NXP internal document […] [Doc ID 1453-9284].

(300) Therefore, the Commission considers that, while the market for the supply of V2X chips is nascent, its relevance in the semiconductor industry for automotive application would significantly increase over the next years.

B) Qualcomm and NXP are not close competitors and alternative operators will remain active

(301) First, while the majority of respondents to the market investigation indicate that both Qualcomm and NXP are active and competing on the market for automotive V2X chips, the market investigation shows that Qualcomm is mainly focused on cellular V2X, which relies on Qualcomm's cellular technology while NXP provides only non-cellular V2X, based on the 802.11.p technology. Respondents indicate that there are differences among the solutions offered by the different suppliers, in particular in terms of security and performance. NXP and Autotalks are viewed as the only two suppliers with a complete V2X offering, including the V2X 802.11.p radio, the secure element and the application processor .

(302) Second, most of the respondents to the phase II market investigation do not believe that Qualcomm and NXP closely compete in relation to the development, manufacturing and sale of non-cellular V2X radio and consider Autotalks as the closest competitor to NXP. This has been further confirmed by the internal documents submitted by the Parties.

(303) NXP's documents confirm that Autotalks […] . […][…] . […] .

(304) Qualcomm also considers […]. Moreover, Qualcomm notes that […] .

(305) Third, most of the respondents to the phase II market investigation consider that a sufficient number of alternative providers of non-cellular V2X chips would remain available post-Transaction. Winstron, Nexty and Commsigina indicate Autotalks,

See responses to questions 41 and 42 of Q6 – automotive customers, responses to questions 40 and 41 of Q5 – automotive competitors. See also responses to question 19 of Q16– automotive customers, responses to questions 15 of Q15 – automotive competitors.

Minutes of the conference call with Autotalks of 19 July 2017. [Doc ID 3264].

See responses to question 20 of Q16– automotive customers, responses to questions 16 of Q15 – automotive competitors.

See responses to question 19 of Q16– automotive customers, responses to questions 15 of Q15 – automotive competitors.

NXP internal documents, “[…]”, dated 18 June 2015 [Filename DOC-000009829 msg] [Doc ID 1454-9829]. See also, “[…]”, dated 23 July 2015 [Filename DOC-000021546 msg] [Doc ID 1454-21546].

NXP internal documents, “[…]”, dated 8 December 2016 [Filename DOC-000005377 msg] [Doc ID 1453-5377].

NXP internal documents, “[…]”, dated 28 November 2016 [Filename DOC-000056281 msg] [Doc ID 1452-56281].

NXP internal documents, “[…]”, dated 28 November 2016 [Filename DOC-000056281 msg] [Doc ID 1452-56281].

Qualcomm internal document, “[…]”, dated 8 March 2016 [Filename […].pptx] [Doc ID 1334-6849]. See also, “ […]”, dated 4 September 2016 [Filename QCRIVEREU_RFI20_1347360.pdf] [Doc ID 2511-35922].

Qualcomm internal document, “[…]”, dated 22 September 2016 [Filename […].msg] [Doc ID 2387-54314].

(306) In internal correspondence, Qualcomm's employees note that […] has become active in DSRC [non-cellular V2X] in the standards forums" and that […] entry may be expected in the near future. Moreover, also […] and […] are considered as actively competing for some specific customers.

C) Market entry

(308) With regard to potential new entry in the market, the phase I market investigation indicates that new suppliers should enter the market over the coming years. However, based on those results, it further appears that market entry would require significant time and investment. Therefore, in the Article 6(1)(c) decision, the Commission noted that it would further investigate whether addition players would start developing V2X chips based on non-cellular technology over the next years and whether the development of such technology would require significant time and investment. In its response to the Article 6(1)(c) decision, the Notifying Party notes

Renesas publicly announced the launch of non-cellular V2X solutions last year. See "Renesas Electronics Delivers Vehicle-to-Vehicle and Vehicle-to-Infrastructure Communication System Solutions for the Autonomous-Driving Era", dated 6 October 2016, available at: https://www.renesas.com/en-us/about/press-center/news/2016/news20161006 html [Doc ID 3267].

Redpine publicly announced the launch of non-cellular V2X solutions designed last year. See "Redpine Signals Launches Industry’s First Multi-Protocol Wireless Solution for the Connected Car (V2X) Market", dated 6 June 2016, available at: http://www.redpinesignals.com/News_&_Events/PressReleases/Redpine_Signals_LaunchesIndustrys_F irst_Multi-Protocol_Wireless_Solution_for_the_Connected_Car_(V2X)_Market.php [Doc ID 3268].

Marvell recently publicly announced the launch of its non-cellular V2X solution. See "Marvell Introduces Industry’s First Wi-Fi, Bluetooth 5 and 802.11p Combo Solutions for Vehicle-to-Everything (V2X) and In-Vehicle Infotainment (IVI)", dated 13 June 2017, available at: https://www.multivu.com/players/English/8119751-marvell-88w8987xa-automotive-wireless-combo-solutions/ [Doc ID 3269].

(309) First, as confirmed by the phase II market investigation, R&D investment to develop non-cellular V2X is not relatively significant. Qualcomm and NXP allocate each year respectively USD […] and […] million on non-cellular V2X corresponding to […]% and […]% of annual R&D spending.

(310) Second, some respondents to the market investigation highlight how entry would be easier for suppliers already active in the supply of Wi-Fi chips based on 802.11.a technology since the relevant know-how would facilitate entry. ON Semiconductor notes that the R&D investment should be minimal for suppliers that have existing 802.11.a technology and experience in the automotive industry. Marvell considers that any Wi-Fi vendor would potentially be an entrant on the non-cellular V2X market .

(311) Therefore, the Commission concludes that barriers to entry the market for the development of non-cellular V2X are not significant and that other suppliers, in particular other automotive suppliers of Wi-Fi chips, might enter the market over the coming years.

D) Qualcomm will continue to invest on both cellular and non-cellular V2X technologies, after the Transaction

(312) In the Article 6(1)(c) decision, the Commission also raised the question as to whether the Transaction would harm the future development of V2X chips based on non-cellular technology. In particular, post-Transaction Qualcomm might be incentivised to focus its investment on cellular V2X technology, where it can leverage its cellular patent portfolio, and remove NXP’s non-cellular V2X market leading technology from the market. As a result of this strategy, Qualcomm would favour the development of cellular V2X chips and delay the deployment of non-cellular V2X chips in the market .

(313) The Commission does not consider that Qualcomm will have an incentive to implement such a strategy for the following reasons.

(314) First, should Qualcomm decide to remove NXP’s offering from the market, it would mainly benefit NXP’s competitors, such as Autotalks, who will profit from the increase in demand by automotive customers which are already requiring non-cellular V2X chips. As described in recital (303), Autotalks is NXP’s closest competitor and it has a comparable product offering which could be able to attract the additional demand diverted from the Parties' customers.

(315) Second, Qualcomm expects that both technologies may coexist in the near future. Such coexistence would enable the support of the technology evolution to 5G while maintaining compatibility between different generations of V2X chips . Qualcomm's future automotive connectivity chipset will include V2X capabilities by including a non-cellular V2X chip. The coexistence of both technologies was also foreseen by NXP, which absent the Transaction, was […] by developing a product which will be "[…]" .

(316) Third, there is no evidence in the Commission file that Qualcomm will decide to stop supplying non-cellular V2X chips. To the contrary, Qualcomm's revenue synergy indicates that Qualcomm expects to achieve significant revenues in the non-cellular V2X market over the next years.

(317) Finally, internal documents further indicate that the deployment of non-cellular V2X technology is driven by regulation and governmental decisions .

6.2.5.3. Conclusion

69

Figure 2: EU OEM V2X feature penetration

[…]

Source: NXP internal document […] [Doc ID 1454-874].

(318) The Commission therefore concludes that the Transaction would not raise competition concerns as to its compatibility with the internal market on the market for the manufacture and supply of automotive chips based on non-cellular V2X technology since the Parties are not the closest competitors, alternatives would remain available and new providers are expected to enter the market over the next years.

6.3. Semiconductors for IoT applications

6.3.1. Introduction

(319) As described in section 5.2.2, both Qualcomm and NXP are active in the manufacture and supply of semiconductors for IoT applications. With regard to the segmentation by semiconductor type, the Transaction gives rise to one horizontally affected market, where the Parties’ combined market share is above 20%: the market for the supply of Bluetooth connectivity chips.

(320) With regard to the segmentation by field of application, the Transaction does not give rise to any horizontally affected markets, as the Parties' combined market share is below 20% in each of those group products.

See “With the aim of increasing safety in road traffic, Volkswagen will enable vehicles to communicate with each other as from 2019”, available at: https://www.volkswagenag.com/en/news/2017/06/pwlan html

Qualcomm internal document, “[…]”, [Filename […].xlsx] [Doc ID 2386-68199].

Qualcomm internal document, “[…]”, dated 14 November 2016, [Filename […].pptx] [Doc ID 2511-34396].

NXP internal document, “[…]”, dated 1 April 2016 [Filename […].msg] [Doc ID 1452-44643]. See also, NXP internal document, “[…]” [Filename […].pptx] [Doc ID 1454-13929].

NXP internal document, “[…]”, dated May 2017, slide 10 [Filename […].pptx] [Doc ID 1453-4115].

(318) […]. See Qualcomm internal document, “[…]”, dated 24 February 2016 [Filename […].docx] [Doc ID 2511-35422].

70

for the manufacture and supply of Bluetooth chips for IoT applications and any potential narrower markets.

6.4. Mobile Audio

6.4.1. Introduction

(328) As described in section 5.2.3.3, both Qualcomm and NXP are active in the manufacture and supply of speech enhancement software and smart amplifier chips. Following the market investigation results, the Commission considers that speech enhancement software and smart amplifier chips are two separate product markets. On both markets, the Transaction gives rise to horizontally affected markets as the Parties' combined market share is above 20%.

6.4.2. Speech enhancement software

6.4.2.1. Notifying Party's views

(329) The Notifying Party argues that, on the market segment for speech enhancement software, product offerings by Qualcomm and NXP are different. While Qualcomm's solutions ("Fluence") are supplied to its baseband chipset customers as an integrated product, NXP's solutions ("LifeVibes") are only sold on a standalone basis.

(330) The Notifying Party indicates that there are various suppliers of speech enhancement solutions including companies (i) providing standalone solutions such as ForteMedia, Knowles/Audience, Cirrus Logic, Philips, Sensory, MightWorks, and Nuance; and, (ii) companies supplying integrated solutions, such as Apple, MediaTek, Microsoft Cortana, Google, and Xiaomi/Pine Cone. Artesis. Concerning the closeness of competition, the Notifying Party argues that there is no competition between "Fluence" and "LifeVibes" due their differences in nature (integrated vs. non-integrated solution), costs, performance, and strategic importance.

(331) Moreover, the Notifying Party indicated that competitors such as Cirrus and MediaTek should be considered as new entrants which would demonstrate the lack of competition concerns arising on the market segment for speech enhancement software.

6.4.2.2. The results of the market investigation and the Commission's assessment

(332) Qualcomm and NXP have a combined market share above 20% in speech enhancement software.

73

(336) First, other competitors would remain in the market post-Transaction including established market players, such as Apple, with a market share of [20-30]%. The presence of alternative providers post-Transaction has been confirmed by the market investigation. The majority of respondents confirmed that other manufacturers and suppliers would remain active post-Transaction .

(337) Second, NXP’s lower market share (approximately [5-10]%, [0-5] percentage points less than the market share presented in Table 6 above on the overall market for the supply of speech enhancement software) indicates that customers of Qualcomm’s baseband chipsets do not particularly value NXP’s offering. The majority of competitors also believe that NXP products are not superior to the ones of its competitors. Knowles indicates that the performance of the NXP product is not as good as Knowles' or Fortemedia's. In addition, customers such as Microsoft would continue to consider alternatives even if Qualcomm were to bundle NXP's speech enhancement software with its baseband chipsets .

(338) Moreover, most of market respondents do not consider barriers to enter in this market to be high. Knowles and MediaTek believe that the barriers to entry are moderate. Moreover, MediaTek's entry in the market indicates that alternative suppliers of baseband chipsets are able to enter the market in the short term.

(339) Finally, most of the respondents to the market investigation consider that the Transaction would not have an impact on the market for the supply of speech enhancement software . For the sake of completeness, one respondent to the market investigation noted that the Transaction could have a negative impact in the Mobile Audio markets since Qualcomm would be able to bundle NXP's speech enhancement software and smart amplifiers with its baseband chipsets . The Commission considers this concern as not merger-specific since Qualcomm already offers Mobile Audio products in the market and, therefore, it could already engage in such conduct pre-Transaction. The Transaction does not change Qualcomm's incentive to engage in such strategy since as described in recital (337) and (348), NXP does not hold a significant presence in those markets, several alternatives would remain available and NXP's offering does not have superior functionalities compared to its competitors.

(340) The Commission concludes that the Transaction would not raise competition concerns as to its compatibility with the internal market on the market for the supply of speech enhancement software and the narrower potential market for the supply of speech enhancement software on devices equipped with a Qualcomm baseband chipset.

See responses to question 85.2 of Q1 – Device OEMs, responses to question 18.2 of Q4 – Mobile Audio competitors.

See responses to question 84.2 of Q1 – Device OEMs, responses to question 15.2 of Q4 Q4 – Mobile Audio competitors.

See responses to question 89 of Q1 – Device OEMs.

See responses to question 16.1 of Q4 – Mobile Audio competitors.

See responses to question 94.4 of Q1 – Device OEMs, responses to question 35.2 of Q4 – Mobile Audio competitors.

(346) Table 9, the Parties would have a combined market share of [20-30]%. However, on this market segment, Qualcomm has a market share of less than [0-5]% by value. When considering shares for the supply of smart amplifiers on devices equipped with a Qualcomm baseband chipset, the Parties would have a combined market share of [20-30]%. However, the increment brought by the Transaction would be minor (approximately [0-5]%).

(347) Moreover, while the Parties have a combined market share above 20%, other competitors would remain in the market post-Transaction including established market players such as Apple, with a market share of [20-30]%. The presence of alternative providers post-Transaction has been further confirmed by the phase I market investigation. The large majority of customers confirm that other manufacturers and suppliers would remain active post-Transaction. Samsung indicates that existing suppliers would expand capacity if the merged entity were to stop supplying its products post-Transaction. One competitor, Xiaomi/PineCone, believes that CirrusLogic, Maxim, Texas Instruments would be able to offer alternatives to customers post-Transaction . Moreover, market respondents do not indicate high barriers to entry in this market segment .

(348) In addition, respondents to the phase I market investigation consider that NXP products are not superior to the ones of its competitors. One customer, HTC, indicates that products from other suppliers compete effectively with NXP solutions. MediaTek confirms that other providers, such as Texas Instruments, offer very good solutions .

(349) Finally, most of the respondents to the phase I market investigation consider that the Transaction would not have an impact on the market for the supply of smart amplifiers. In particular, although some customers, such as TCL mentioned the risk of product choices elimination due to bundling practices, others such as Microsoft believe it will be able to obtain supply from other competitors post-Transaction .

(350) The Commission considers that the Transaction would not raise competition concerns as to its compatibility with the internal market on the market for the manufacture and supply of smart amplifier chips and the narrower potential market for the supply of smart amplifier chips on devices equipped with a Qualcomm baseband chipset.

7. COMPETITIVE ASSESSMENT – CONGLOMERATE NON-COORDINATED EFFECTS

(351) Under Article 2(2) and (3) of the Merger Regulation, the Commission must assess whether a proposed concentration would significantly impede effective competition in the internal market or in a substantial part of it, in particular through the creation or strengthening of a dominant position.

(352) In this respect, a merger may entail horizontal or non-horizontal effects, or both. Non-horizontal effects deriving from a concentration may be vertical or conglomerate.

(353) Vertical effects are those deriving from a concentration where the undertakings concerned are active on different or multiple levels of the supply chain. Conglomerate effects are those stemming from a concentration where the undertakings are active on closely related markets (for example, suppliers of complementary products or products that belong to the same product range). In case of a concentration giving rise to conglomerate effects, the Commission will assess the concentration in accordance to the framework set by the Non-Horizontal Merger Guidelines .

(354) Qualcomm is active in the manufacturing and sale of baseband chipsets to device OEMs, whereas NXP supplies NFC and SE chips to device OEMs. NXP also provides the MIFARE technology to device OEMs and licenses MIFARE to SE and SE OS manufacturers.

(355) Based on the information provided by the Notifying Party, the Parties have a common pool of customers: […] , […] ; […].

(356) Moreover, the results of the phase I market investigation indicate that device OEMs generally purchase both baseband chipsets and NFC and SE chips .

(357) Therefore, Qualcomm's baseband chipsets and NXP's NFC/SE products can be considered to be complementary within the meaning of paragraph 91 of the Non-Horizontal Merger Guidelines.

(358) Accordingly, in the following the Commission will examine whether the Transaction may give rise to conglomerate effects in relation to these products.

(359) As described in sections 6.4.2 and 6.4.3 above, Qualcomm and NXP are also active in the supply of smart amplifier chips and speech enhancement software to device OEMs. However, as explained in those sections, the Commission considers that the Transaction does not raise competition concerns as to its compatibility with the internal market on those markets. For the sake of completeness, the Commission notes that one respondent to the phase I market investigation noted that the Transaction could have a negative impact on those markets since Qualcomm could be able to offer NXP's speech enhancement software and smart amplifiers with its baseband chipsets. The Commission however considers this concern as not merger-specific since Qualcomm already offers Mobile Audio products in the market and, therefore, it could already engage in such conduct pre-Transaction. Moreover, the Transaction does not change Qualcomm's incentive to engage in such strategy since as described in Sections 6.4.2 and 6.4.3 above, NXP does not hold a significant presence in those markets, several alternatives would remain available and NXP's offering does not have superior functionalities compared to its competitors.

Guidelines on the assessment of non-horizontal mergers under the Council Regulation on the control of concentrations between undertakings ("Non-Horizontal Merger Guidelines"), OJ C 265, 18.10.2008, paragraph 7.

According to Qualcomm, Apple is an indirect Qualcomm baseband chipset customer in that its devices are made by contract manufacturers, which source chipsets from suppliers like Qualcomm and pass-on their costs to Apple.

Form CO, paragraph 918, Tables 4.4 and 4.5 [DOC ID 326].

See responses to question 3 of Q1 – Questionnaire to device OEMs.

See responses to question 34 of Q4 – Mobile Audio competitors.

(360) Section 7.1 below summarises the analytical framework applicable to conglomerate relationships. Section 7.2 sets out the Parties' and their competitors' market shares for the products relevant for the conglomerate assessment, namely baseband chipsets, NFC and SE chips. Section 7.3 assesses whether the merged entity would hold market power with respect to: LTE baseband chipsets (7.3.1); NFC chips, SE chips and combined NFC/SE solutions (7.3.2); transit service technologies (7.3.3); and IP (7.3.4). In Section 7.4, the Commission examines conglomerate effects in relation to Qualcomm's baseband chipsets, NXP's NFC and SE chips and MIFARE. Section 7.5 analyses conglomerate issues related to the changes the Transaction may bring to NXP's licensing of its IP for NFC technology. Finally, 7.6 Section discusses conglomerate concerns related to Qualcomm's baseband chipsets, NXP's NFC and SE chips as well as the licensing of the merged entity's IP.

7.1. Analytical framework

(361) Conglomerate mergers consist of mergers between companies that are active in closely related markets, for instance suppliers of complementary products or of products which belong to a range of products that is generally purchased by the same set of customers for the same end use .

(362) According to the Non-Horizontal Merger Guidelines, in most circumstances, conglomerate mergers do not lead to any competition problems .

(363) However, foreclosure effects may arise when the combination of products in related markets may confer on the merged entity the ability and incentive to leverage a strong market position from one market to another closely related market by means of tying or bundling or other exclusionary practices .

(364) The Non-Horizontal Merger Guidelines distinguish between bundling, which usually refers to the way products are offered and priced by the merged entity and tying, usually referring to situations where customers that purchase one good (the tying good) are required to also purchase another good from the producer (the tied good) .

(365) Within bundling practices, the distinction is also made between pure bundling and mixed bundling. In the case of pure bundling the products are only sold jointly in fixed proportions. With mixed bundling the products are also available separately, but the sum of the stand-alone prices is higher than the bundled price .

(366) Tying can take place on a technical or contractual basis. For instance, technical tying occurs when the tying product is designed in such a way that it only works with the tied product (and not with the alternatives offered by competitors). While tying and bundling have often no anticompetitive consequences, in certain circumstances such practices may lead to a reduction in actual or potential competitors' ability or incentive to compete. This may reduce the competitive pressure on the merged entity allowing it to increase prices .

See Non-Horizontal Merger Guidelines, paragraph 91.

See Non-Horizontal Merger Guidelines, paragraph 92.

See Non-Horizontal Merger Guidelines, paragraph 93.

See Non-Horizontal Merger Guidelines, paragraph 97.

See Non-Horizontal Merger Guidelines, paragraph 96.

(371) Table 11 and

(377) In its response to Article 6.1(c) decision, the Notifying Party emphasizes that its declining market share provides strong evidence that it faces aggressive competition. Furthermore, it contests the reliability of a third party report's forecast of its market share in 2020. In addition, it points out that its rivals exert a strong competitive constraint, as indicated by the drop in its market share. Finally, it contests the importance of its position as supplier of CDMA chipsets because device OEMs choose to release the same model across all networks (but are not obliged to do so), CDMA is a disappearing technology and, in any event, CDMA chipsets represent a small portion of OEMs' chipset demand.

(378) At the outset, the Commission recalls that, according to the Non-Horizontal Merger Guidelines, market shares provide useful first indications of the market power and the competitive importance of both the merging parties and their competitors .

(379) Against this background, the Commission first notes that, based on the information presented in Section 7.2.1 above, Qualcomm has high market shares: based on the 2016 figures, Qualcomm's market share would be of [60-70]% by revenues with respect to LTE baseband chipsets Qualcomm's market share would remain high even when considering overall baseband chipsets ([50-60]% by revenues when considering the overall possible market, [50-60]% when excluding captive sales).

(380) Third party industry reports also confirm this. For instance, the Linley Group report "Mobile Semiconductor Market Share Forecast 2015-2020" (which takes into account captive sales by Samsung LSI and HiSilicon) attributes to Qualcomm a 2016 market share of [50-60]% by revenues, well ahead of MediaTek. While the Linley Group report forecasts […], it still attributes to Qualcomm a market share in […] for LTE baseband processors of [40-50]% by unit shipments and of [50-60]% by revenues and states that the company […]. Qualcomm has contested the reliability of such forecast, without however offering any further explanation as to reasons why such figure should be disregarded or offering alternative predictions about the evolution or level of its market shares in the future.

(381) In any event, Qualcomm's 2016 market share in revenue terms is indicated as being close to [50-60]% with regard to all baseband chipsets, and above [50-60]% when considering only LTE baseband chipsets, also by other third party industry reports.

(382) As such, despite exhibiting a relative decline, Qualcomm's market share is still substantial and remains above [50-60]%. This provides a first indication that Qualcomm enjoys market power at the very least with respect to LTE baseband chipsets.

(383) The Commission also notes that respondents to the phase I market investigation generally emphasised that Qualcomm has a strong market position in baseband chipsets. Qualcomm was indicated by device OEM customers and baseband competitors as the leading provider for baseband chipsets in terms of technology, quality, offering breadth, and integration. Providers of NFC technology also commented that Qualcomm is a strong player as regards baseband chipsets .

(384) Furthermore, the phase I and phase II market investigation indicated that, contrary to the Notifying Party's assertions, there are no alternative providers of baseband chipsets that would be capable of constraining Qualcomm's market power. Respondents generally indicated that providers such as Spreadtrum, MediaTek and Intel do not exercise a strong competitive constraint on Qualcomm. Samsung LSI and HiSilicon were also not viewed as viable alternative suppliers, as they are focused on non-merchant sales for their parent companies, Huawei and Samsung .

(385) As regards Spreadtrum, the Commission notes that its baseband production mainly […] and […].

(386) MediaTek began shipping LTE baseband chipsets only in 2014 and, while growing, it is still behind Qualcomm in terms of market share. In that regard, Intel commented that MediaTek is a supplier of "mainstream baseband chipsets (i.e., non-premium LTE baseband chipsets)". Indeed, MediaTek's offering appears to be focused on low to mid-range segments of LTE baseband chipsets and on sales in China .

(387) Research by ABI research from 17 February 2016 indicate that, while Qualcomm is […] .

(388) Industry reports also note that […] .

shares. Therefore, should […]'s baseband chipset sales be excluded, Qualcomm's market share would likely be even higher.

See responses to questions 18-22 of Q1 – Questionnaire to device OEMs; responses to questions 18-21 of Q2 – Questionnaire to baseband competitors.

See responses to question 22 of Q1 – Questionnaire to device OEMs; responses to question 19 of Q2 – Questionnaire to baseband competitors; LG's response to question 50 of Q9 – Questionnaire to device OEMs, [Doc ID: 2043].

Responses to question 22 of Q3 – Questionnaire to NFC competitors.

Samsung LSI only sells baseband chipsets to a Chinese OEM (see Samsung's response to question 5 of Commission RFI 28, dated 22 July 2017, [Doc ID: 2671]).

See responses to questions 18-22 of Q1 – Questionnaire to device OEMs; responses to questions 18-21 of Q2 – Questionnaire to baseband competitors.

The Linley Group: "Mobile Semiconductor Market Share Forecast 2015-2020", pages 40, and 42; Form CO, Annex 4.16 [Doc ID: 327].

See Intel's response to question 18 of Q2 – Questionnaire to baseband competitors, [Doc ID: 767].

See Strategy Analytics' September 2016 baseband market share tracker, which indicated that "[…]" (Annex 4.10 to the Form CO).

ABI Research Forecasts LTE Carrier Aggregation to Power 61% of Smartphones Shipped in 2020, from https://www.abiresearch.com/press/abi-research-forecasts-lte-carrier-aggregation-pow/, consulted on 25 October 2017 (Doc ID 3270). LTE CA stands for LTE-Advanced chipsets that form part of the (high-end) LTE chipsets market.

The Linley Group: "Mobile Semiconductor Market Share Forecast 2015-2020", page 41; Form CO, Annex 4.16 [Doc ID: 327].

(389) With respect to Intel, the Commission notes that, to date, it supplies LTE standalone baseband chipsets only to one customer, Apple. Intel's entry in the market is recent, and it is yet to be seen whether Intel will develop an integrated baseband chipset offering.

(390) In addition, in line with the results of the phase I and phase II market investigation, the Commission considers that the worldwide market for LTE chipsets is characterised by the existence of a number of barriers to entry and expansion, including in relation to (i) R&D activities; (ii) certification and relationships with OEMs and MNOs; and (iii) the importance for suppliers to supply chipsets supporting a variety of standards.

R&D activities

(391) A new supplier of LTE chipsets needs to undertake significant initial investments in R&D activities related to the design of LTE chipsets in order to launch its first product on the market.

(392) Indeed, the Notifying Party confirms that R&D plays an important role in the whole semiconductor industry, with new and improved products continuously being introduced.

(393) The Notifying Party itself has invested heavily into R&D: it has spent over USD 43 billion since its founding in 1985, and USD 5.5 billion, USD 5.5 billion and USD 5.2 billion in 2014, 2015 and 2016, respectively. It further acknowledges that past R&D expenditure benefits current and future developments, thus having a compounding effect.

(394) Competing suppliers of LTE baseband chipsets also indicated that they invested heavily into R&D.

Certification and relationships with OEMs and MNOs

(395) Baseband chipsets need to be certified by MNOs on their networks and by OEMs in their devices. The process, which takes at least between 6 to 12 months, constitutes a barrier to entry because OEMs and MNOs are often reluctant to test offerings from new baseband chipset suppliers, in light of the significant time and resources required for such activity. This heightens the risk for a new entrant of committing sunk costs to the development of a new baseband chipset.

See Intel's response to question 3 of Q2 – Questionnaire to baseband competitors, [Doc ID: 767]. In 2016, "[…]" (The Linley Group: "Mobile Semiconductor Market Share Forecast 2015-2020", page 33; Form CO, Annex 4.16 [Doc ID: 327]).

See, for example, replies to question 30 of Q2 – Questionnaire to baseband competitors.

(396) In addition, established suppliers of baseband chipsets benefit from time savings in the certification by OEMs and MNOs due to similarities between older and newer chipset models of the same supplier, since testing can focus on new features.

(397) Furthermore, competitors point out that, in order to develop baseband chipsets, a potential supplier needs to cooperate with a number of players, including equipment manufacturers and MNOs "to understand how the standards are implemented at the end product level" and align its offerings with their equipment and networks. Engagement with handset manufacturers is considered "critical" because they provide feedback from deployment of chipsets in a diverse range of real world conditions. The baseband chipset manufacturers use this feedback to improve the reliability and performance of their chipsets.

Importance for suppliers to supply chipsets supporting a variety of standards

(398) OEMs often sell the same devices throughout the world and they expect that suppliers are able to provide baseband chipsets that support a variety of standards used across all geographic areas. In particular, two of the largest carriers in the USA (Verizon and Sprint) as well as several Chinese carriers own mobile networks based on the CDMA standard.

(399) Accordingly, it is important that a supplier is able to supply LTE baseband chipsets that are also compliant with CDMA. This constitutes a barrier to entry since the inability to provide a uniform product globally may deter OEMs from sourcing baseband chipsets from suppliers not offering baseband chipsets compliant with CDMA, in light of the costs associated with developing separate products (for CDMA and non-CDMA baseband chipsets).

(400) In this respect, until 2015, there were only two suppliers active in the production of CDMA baseband chipsets, Qualcomm and VIA Telecom. In 2015, Intel bought the CDMA assets of Via Telecom and has since integrated that technology to begin to develop a CDMA-enabled multimode LTE baseband chipset, which should reach the market only in 2018. In 2015, MediaTek also started supplying CDMA chipsets. However, it addresses lower end segments of the market and its products do not constitute a viable alternative for OEMs requiring more premium features. Therefore, Qualcomm is in practice the only viable supplier of baseband chipsets that are compatible with any CDMA standard.

See Intel's response to question 24 of Q10 – Questionnaire to baseband competitors, [Doc ID: 2307].

See Samsung's response to question 1 of Commission RFI 28, dated 22 July 2017, [Doc ID: 2671].

See Intel's response to question 25 of Q10 – Questionnaire to baseband competitors, [Doc ID: 2307].

See Intel's response to question 25 of Q10 – Questionnaire to baseband competitors, [Doc ID: 2307].

(401) The fact that major device OEMs (such as Samsung) need to continue sourcing from Qualcomm LTE chipsets compliant with the CDMA standard hampers their attempts to switch away from Qualcomm.

Conclusion

(402) Based on the above, the Commission considers that Qualcomm holds a dominant position in the worldwide market for LTE baseband chipsets.

(403) Additionally, the Commission notes that several respondents to the Commission's questionnaires raised the issue of Qualcomm's so-called "grant-back network" constituting a barrier to entry. Given that the Commission concludes that Qualcomm holds a dominant position in the worldwide market for LTE baseband chipsets, there is no need to conclude on whether Qualcomm's so-called "grant-back network" constitutes an additional barrier to entry.

(404) Therefore, after the Transaction, the merged entity would hold a dominant position in the worldwide market for LTE baseband chipsets.

7.3.2. Market power of the merged entity with regard to NFC, SE chips and combined NFC/SE solutions

7.3.2.1. Notifying Party's views

(405) The Notifying Party acknowledges that NXP holds a high share in the supply of NFC and SE chips. However, such market shares are not indicative of market power for several reasons. In particular, NXP's position is largely attributable to sales to Apple and Samsung and, in light of its […] share of NXP's sales, Apple is aware of its importance for NXP's NFC and SE business. Furthermore, NXP faces intense competition from other credible players […]. In addition, the market for NFC and SE is nascent, which precludes the finding of market power.

(406) The Notifying Party further argues that these products are not essential to the functionality of a mobile device and face competition from competing technologies, such as BTLE, QR codes, MST, "HotKnot" technology and WMC technology (for NFC), HCE, TEE and TPM (for SE).

(407) In its response to the Article 6.1.(c) decision, the Notifying Party reiterates that NFC and SE do not have a "must have" nature, as evidenced by the fact that most of mobile devices are currently sold without such technologies. Furthermore, it points out that the attach rate of NFC and SE for mid-tier mobile devices in 2020 (respectively [40-50]% and [5-10]%) provide a good indication that competing technologies (for example, HCE) exert competitive pressure on either or both product segments.

7.3.2.2. The results of the market investigation and the Commission's assessment

(408) Based on the results of the phase I and phase II market investigation and on the review of the Parties’ internal documents, the Commission considers that post-Transaction the merged entity would have a certain degree of market power as regards NFC and SE chips, as well as the combined NFC/SE solutions.

See Samsung's response to question 1 of Commission's RFI 28, dated 22 July 2017, [Doc ID: 2671].

The "grant-back network" would be premised on the fact that, when Qualcomm agrees cross-licences with other holders of IPR in the UMTS and LTE standards, Qualcomm obtains the right of pass-through of the other party's IPR to Qualcomm's chipset customers.

Market shares

(409) NXP's market share is particularly high with regard to each of NFC and SE chips ([70-80]% and [60-70]% by revenue, respectively).

(410) With regard to NFC chips, the Linley Group report indicates that in 2016 NXP's market share was of [60-70]% by shipments and [60-70]% by revenues, well ahead of the closest competitor, […]. It should be noted that the Linley Group report includes in the NFC market the sales of Broadcom, which the Commission understands has exited the NFC space, and of Samsung LSI, which mostly sells internally to Samsung. Therefore, NXP's position in the NFC market is likely even stronger.

(411) However, NXP's high market shares may overestimate its market power since the phase II market investigation has revealed that NXP's sales are dependent on two large customers […], which together account for more than [80-90]% of NXP's NFC and SE sales by volume. A decision by these customers to switch away from NXP and adopt a "mix-and-match" approach would dramatically reduce NXP's market share in NFC/SE.

(412) Apple and Samsung already "mix-and-match" NFC chips and SE chips from different suppliers: […]. Even if it continues to procure a share of its total demand for NFC/SE solutions from NXP, Samsung has increased its reliance on its in-house capacity of NFC and SE chips, mixed-and-matched with an SE OS from a third party player.

(413) Moreover, the fact that Apple represents a substantial part of NXP's NFC/SE sales is in the public domain. Therefore, Apple is aware of the relevance of its purchases for NXP's NFC and SE business. In this respect, in its economic submission, […] .

Competitive pressure from suppliers of NFC chips, SE chips and "mix-and-match" solutions

(414) In the phase I market investigation, NXP was indicated by device OEM customers as the strongest market player in terms of all the relevant components of NFC technology (NFC chip, SE chip and SE OS). NXP was depicted as the market leader, with the most widely adopted and innovative solution on the market. The fact that NXP offers all three relevant components also was indicated as a point of strength. Competitors also indicated NXP as the leading player and highlighted the strength of NXP’s complete offering. Some respondents also indicated that MediaTek is not active in the market, if not completely exiting it. Providers of single components were indicated as generally being less innovative or, such as Samsung LSI, mainly producing for its internal demand.

The Linley Group: "Mobile Semiconductor Market Share Forecast 2015-2020", pages 101 – 103; Form CO, Annex 4.16 [Doc ID: 327].

See non-confidential minutes of conference call with Broadcom on 11 July 2017.[DOC ID 1961]

See reply to Article 6.1.(c) decision, paragraphs 43-45.

See non-confidential version of the Notifying Party's presentation of 29 June 2017. [Doc ID: 1320]

See […] submission, "Economic model of Qualcomm’s licensing practices", dated 1 September 2015, p. 13. [Doc ID: 2564]

See responses to questions 19-21 of Q1 – Questionnaire to device OEMs.

Responses to questions 19-21 of Q3 – Questionnaire to NFC competitors.

(415) However, the phase II market investigation provided indications that the "mix-and-match" solutions based on components from different suppliers do exercise a competitive pressure on NXP's NFC/SE solution.

(416) In particular, the majority of device OEMs as well as NXP's competitors have indicated that there are viable alternative to NXP for the supply of each of NFC chips (including STMicroelectronics, Samsung, Sony, HiSilicon, and Mediatek), SE chips (including STMicroelectronics, Samsung, and Infineon) and SE OS (including Gemalto, G&D, and Oberthur). Moreover, contrary to the opinion expressed by some respondents in the phase I investigation, MediaTek appears to be active in the market, by way of example having provided the NFC chip for a handset by Nokia released in September 2017.

See responses to questions 28 and 29 of Q9 – Questionnaire to device OEMs.

See report from Tech Insights entitled "Nokia 3 TA-1032 ID253167-AKd" (attached to the Notifying Party's email dated 15 September 2017). [Doc ID: 3331]

(417) In addition, the majority of respondents consider that the resulting "mix-and-match" solutions constitute viable alternatives to NXP's combined solution and point out that manufacturers such as STMicroelectronics, Samsung LSI or Infineon provide these offerings.

(418) Moreover, some device OEMs have indicated that they multi-source for their demand of NFC and SE chips, usually for reasons of maintaining competitiveness in terms of quality and price, thus providing further support to the notion that NXP is faced with competition from other suppliers.

(419) Indeed, some device OEMs stated that in their negotiations with NXP they have been able to threaten to switch to an alternative provider and extract better terms. According to a competitor, switching providers is common since, for each new device generation, device OEMs define their requirements and decide which components to integrate in any given product, which "opens, in principle, for each device generation (almost every year) a design-in window at device OEMs".

(420) While the strength of NXP's product offering for NFC and SE is reflected in the Parties' internal documents, these documents also provide NXP's views on the existing competition on the NFC/SE markets.

(421) With respect to Samsung LSI, NXP's internal correspondence show that Samsung's use of a mix-and-match solution […] was seen as a […] and that Samsung LSI […] . Moreover, internal documents also mention increasing competition with […] . Similarly, HiSilicon's […] .

See responses to questions 28 and 29 of Q9 – Questionnaire to device OEMs.

See report from Tech Insights entitled "Nokia 3 TA-1032 ID253167-AKd" (attached to the Notifying Party's email dated 15 September 2017). [Doc ID: 3331]

See responses to questions 28 and 29 of Q9 – Questionnaire to device OEMs.

See responses to question 51 of Q9 – Questionnaire to device OEMs.

See responses to question 58 of Q9 – Questionnaire to device OEMs.

See Infineon's response to question 27 of Q11 –Questionnaire to NXP competitors. [Doc ID: 2102]

(422) In this respect, the phase II market investigation has shown that there is also some degree of potential competition coming from players currently selling outside of Europe. In particular, one supplier of NFC chips active in Japan indicated that it intends to enter the European market.

(423) Finally, as discussed in Section 5.2.3.2 above, the phase II market investigation has shown that alternative technologies, in particular TEE and HCE, may constitute a competitive alternative to embedded SEs and that, therefore, a certain degree of substitutability may, therefore, exist between the various technologies, especially from a demand side.

(424) In light of the above findings, the Commission therefore considers that NXP holds a certain degree of market power in the markets for NFC chips, SE chips and combined NFC/SE solutions.

(425) The merged entity would thus hold a certain degree of market power in the markets for NFC chips, SE chips and combined NFC/SE solutions.

Market power of the merged entity in relation to transit services technologies

(426) In this Section, the Commission assesses whether the merged entity would hold market power within the market for transit service technologies and for the purpose of the deployment of transit services through mobile devices, as the merged entity would control MIFARE, NXP’s proprietary technology for transit services. Other technologies within this market include Calypso, CIPURSE and FeliCa.

Notifying Party's views

(427) The Notifying Party acknowledges that MIFARE is the leading platform in the transit space today: however, the Notifying Party specifies that such is the case for card ticketing services, not for smartphone ticketing services. Moreover, in the Notifying Party’s view, there are many competing alternatives to MIFARE that transit authorities could choose from as they upgrade their systems. Thus, solutions transit authorities may adopt as an alternative or in addition to MIFARE include CIPURSE, Calypso and FeliCa. Transit authorities may also adopt an "open loop" system that allows customers to pay directly from their credit cards or bank accounts as opposed to requiring customers to purchase "closed loop" transportation credit in a separate transaction.

(428) Furthermore, the Notifying Party submits that MIFARE is not a must-have product, particularly for the deployment of transit services through mobile devices. According to the Notifying Party, there is currently no transport ticketing implementation of MIFARE on smartphone devices supplied by […], or other major device OEMs, and only a small share of shipped SE chips are even MIFARE enabled.

(429) Moreover, the Notifying Party argues that there are limited prospects for generalized use of MIFARE on mobile phones because, MIFARE requires a different implementation for each transit system, and transit systems have or are likely to upgrade to EMV (contactless payment using banking card data) instead.

(430) In the reply to the 6(1)(c) decision, as regards market power of MIFARE, the Notifying Party further argues that there is no actual implementation of MIFARE for use on mobile devices, which undermines any finding of relevance or importance of MIFARE for device OEMs. In this respect, the Notifying Party contests that, to the extent that certain mobile devices include a MIFARE-enabled SE chip and are thus MIFARE-enabled, that does not amount to an implementation of MIFARE on the mobile device for actual use in transit services. Further, the Notifying Party submits that MIFARE would be important for device OEMs to the extent that contactless applications for transit are important for mobile devices, which is not the case.

(431) Furthermore, the Notifying Party submits that, while it is true that MIFARE has a [70-80]% market share as regards contactless ticketing, that share relates to smartcards, but not smartphones or mobile devices. Moreover, the Notifying Party explains that there are different types of MIFARE licence (Classic, DESFire, Plus), one of which (Classic) does not require a licence from NXP: a breakdown of MIFARE’s share on the basis of those versions would reduce the actual share of MIFARE. In that regard, the Notifying Party also explains that implementation of MIFARE may not require a licence to MIFARE overall, but only to one type of MIFARE.

(432) Finally, the Notifying Party reiterates that there are existing alternative technologies to MIFARE, such as FeliCa and Calypso, which exert a competitive pressure.

7.3.3.2. The results of the market investigation and the Commission's assessment

(433) Based on the results of the phase I and phase II market investigation and on the review of the Parties’ internal documents, the Commission considers that the merged entity would hold a dominant position in the market for transit service technologies given the importance of MIFARE, which is also important for the deployment of transit services for mobile devices.

(434) First, the Commission notes that NXP explains on its official website that MIFARE is present in 750 cities and estimates that it has a market share of 77%. This share, which is above 50%, provides a first indication that MIFARE today enjoys market power for transit/ticketing/fare collection.

(435) The Commission also notes that MIFARE is particularly prevalent in Europe and in the United States today, with a large installed base in many large cities. To the contrary, alternative transit services have a more limited presence: FeliCa is only used in Japan, whereas Calypso is present in Europe, but only in certain countries and cities.

(436) Moreover, as was noted by respondents to the phase I market investigation, transit authorities seldom change their transit service technology, given that this entail significant difficulties and switching costs. […]. Samsung expressed the same view, stating that "NXP's MIFARE technology already has become entrenched in contactless transit payment market by virtue of the significant hardware investments already made by transit merchants across the world. These hardware investments create a "lock-in" effect which heavily incentivizes the continued use and development of MIFARE technology, and deters the adoption of competing technologies which would require either abandonment of existing hardware or additional parallel investments in hardware compatible with other technologies".

See www.nxp.com/video/:MIFARE –HISTORY.

See response by […] to question 46.1. of Q1 - Questionnaire to device OEMs, [Doc ID: 1009].

See response by Samsung to question 50.1 of Q1 - Questionnaire to device OEMs, [Doc ID: 1092].

(437) Second, during the phase I investigation, contrary to the Notifying Party’s view, the majority of respondents among device OEMs, baseband providers and NFC/SE providers emphasised that, already today, it is important to offer MIFARE-enabled NFC and Se products and that MIFARE’s importance is likely to increase in the next two to three years.

(438) For example, […] .

(439) Most of those respondents also stated that, in their view, MIFARE is already today a “must have” feature for smartphone contactless applications for transit/ticketing/fare collection. Device OEMs counting for a significant part of the market also indicated that they are already purchasing MIFARE-enabled NFC/SE and that they require NFC/SE technology to be MIFARE enabled (this was also confirmed by providers of NFC/SE products).

(440) In that regard, […]. Samsung replied that "MIFARE has become a de facto standard at a minimum in transit ticketing and payment industry. […] Given NXP's dominance in NFC and NFC+SE ICs, which could be compatible with MIFARE, MIFARE will likely achieve even greater penetration, including in the general contactless mobile payment industry". Gemalto pointed out that "[m]any device OEMs require MIFARE enabled NFC/SE.", and Toshiba argued that "MIFARE currently has a high share as a smartphone contactless application on a worldwide basis".

(441) The responses to the phase I market investigation also did not support the argument that other transit technologies, such as FeliCa and Calypso, are valid alternatives to device OEMs, for the purpose of mobile transit services. In that regard, it appears that OEMs do not purchase other technologies such as FeliCa and Calypso as alternatives to MIFARE, but rather as complements (FeliCa for use on smartphones mainly, if not exclusively, in Japan, while Calypso is used in Europe and the US).

(442) […] . […] .

(443) Moreover, when asked to estimate how the installation of MIFARE would develop in the next two to three years as compared to alternative technologies (Calypso, FeliCa), most devices OEMs responding to the phase I market investigation indicated that MIFARE would remain more important. Competing NFC/SE providers were mostly of the view that MIFARE would become more relevant than FeliCa and Calypso .

(444) Third, in the course of its phase II investigation, the Commission further requested the feedback of market participants on the importance of MIFARE.

(445) The phase II market investigation confirmed that MIFARE is the most relevant transit service technology, and that existing alternative technologies, such as Calypso and FeliCa, have limited presence and reach.

(446) In particular, transit authorities responding to the Commission’s phase II market investigation indicated that MIFARE presents advantages compared to other transit technologies, such as reliability, executions speed, implementation and maintenance costs, as well as broad market support. Those respondents also indicated that they are not planning to change to another transit technology different from MIFARE .

(447) In that regard, one market player within the NFC/SE space explained that MIFARE has a high market share not only in terms of installed base, but also in terms of new shipments, pointing to the fact that in Europe, FeliCa is not present, Calypso has a share of 5% in terms of installed base and 1% in terms of shipments, with the rest of Europe being covered by MIFARE (which also counts for 97% of shipments) .

(448) […] .

(449) With respect to the importance of MIFARE for device OEMs, […] and Samsung confirmed that they have MIFARE-enabled SEs installed in their latest mobile devices and require MIFARE as a necessary feature for SE chips. […] . […] . Samsung indicated that its latest devices sold in the European Union include MIFARE Classic and that “Samsung requires MIFARE compatibility in all of Samsung’s latest handset models that have SE” .

(450) Other device OEMs also indicated that they consider MIFARE as an important feature for the selection of SE chips. For instance, LG commented that “MIFARE support can be an important factor when considering SEs” and HTC explained that “the acquired SE must offer MIFARE”. Those devices OEMs that provided a meaningful response to the phase II market investigation also indicated that MIFARE is and will remain important for the purpose of transit services in the EEA .

(451) NFC and SE providers also submitted that, in their experience, device OEMs require SE chips to be MIFARE-enabled. For instance, Gemalto commented that in the last three years of tenders launched by device OEMs for SE chips, all required MIFARE implementation . Gemalto further explained that “[p]roposing a product without MIFARE is perceived as lack of competitiveness and OEMs do not want to the risk of ‘limiting’ potentially their service reach. So not proposing MIFARE is not an option” . STMicroelectronics also commented that “MIFARE is an essential feature required by almost all OEMs” and that “[o]ffering MIFARE allows ST to meet the very minimum level of requirements of our customers. When starting the promotion of any products with our customers, MIFARE is a minimum threshold requirement highlighted by the customers to continue the discussion”. Infineon also stated that “MIFARE is the most prevalent technology in tenders received in recent years” .

See responses to question 48 of Q3 – Questionnaire to NFC competitors.

See responses to question 4 of Q12 – Questionnaire to public transit authorities.

See responses to question 7 of Q12 – Questionnaire to public transit authorities.

Minutes of the conference call with Infineon of 22 June 2017, point 4 [Doc ID 1425].

Minutes of conference call with […] of 26 June 2017 [Doc ID 1356].

(455) The Commission also notes that NXP’s own website provides an illustration of the launch of transit service functionalities through NFC, MIFARE and mobile devices. For instance, in a presentation entitled “MIFARE4MOBILE” from 2015, NXP listed examples of cities where the roll out of NFC mobile ticketing with MIFARE was under way .

(456) Fourth, the Parties’ internal documents indicate that MIFARE is the most relevant technology for the purpose of transit services and mobile transit, and that both Parties attribute growing importance to the development of transit services through mobile devices, which includes MIFARE technology.

also requested as a mandatory requirement by MNOs, see its reply to question 2.2.2. This is also confirmed by Qualcomm’s internal document mentioned in recital (471) below, […].

See Gemalto’s reply to Commission RFI 25 of 4 July 2017, question 2.2.5. See also Gemalto’s reply to question 1.17, stating that MIFARE was a “must have”, [Doc ID: 2152].

See STMicroelectronics’ reply to Commission RFI 30 of 5 July 2017, questions 2.2.1 and 2.2.2 [Doc ID: 2811].

See Infineon’s reply to Commission RFI 29 of 5 July 2017, question 2.2.5 [Doc ID: 2955].

See Gemalto’s reply to Commission RFI 25 of 4 July 2017, questions 2.2.3 and 2.2.4 [Doc ID: 2152]; see G&D’s reply to Commission RFI 26 of 4 July 2017, questions 2.2.3 and 2.2.4 [Doc ID: 1932]; see Infineon’s reply to Commission RFI 29 of 5 July 2017, questions 2.2.3 and 2.2.4 [Doc ID: 2955].

(457) In several internal documents, NXP acknowledges and presents MIFARE’s primacy as transit service technology and its importance for the deployment of mobile transit services. In one presentation, […]. Another NXP document, […] .

(458) NXP reiterates the importance of MIFARE in various other internal documents. […] .

(459) In another internal email […]. This NXP exchange indicates the importance of MIFARE for transit services as well as for device OEMS seeking to deploy mobile transit services.

(460) In another presentation, NXP reiterates the predominance of the MIFARE standard against other transit technologies, as reproduced in Figure 3 below.

Figure 3: MIFARE worldwide presence for e-ticketing

[…]

Source: NXP internal document […], November 2013, slide 7 [DOC ID 2941]

(461) The same document also explains that MIFARE […] .

(462) The Commission further notes that these NXP internal documents emphasise the importance and market presence of MIFARE as an overall technology, regardless of the specific type of MIFARE technology deployed or installed. Thus, contrary to the Notifying Party’s argument that a breakdown of MIFARE by type or generation would dilute MIFARE’s share, MIFARE appears to be considered by NXP itself as the dominant transit service solution, regardless of the specific MIFARE considered. Moreover, as can be seen from the market industry reports presented below in recital (477) and Figure 6, even in case of a breakdown of MIFARE per type of technology, MIFARE would still be the most deployed and installed transit protocol technology.

(463) With respect specifically to the Notifying Party’s argument that it is not necessary for a competitor to hold a licence to MIFARE Classic (and the example that Infineon offers “MIFARE compatible” products for MIFARE Classic), the Commission notes that NXP internal documents discussing the implementation of […]. This NXP internal discussion therefore contradicts the Notifying Party’s argument that it is not necessary to hold a licence from NXP to use MIFARE Classic.

(464) The Commission also notes that NXP’s MIFARE4MOBILE technology enables the use of transit services through NFC and mobile devices regardless of the relevant type of MIFARE employed by a transit system. NXP explains that MIFARE4MOBILE “provides mobile network operators, trusted services managers and service providers with a single, interoperable programming interface to

NXP internal document, presentation for “[…]”, slide 10, attached to NXP email by […] of 24 January 2017, [DOC ID 1456-54560]

NXP internal document, “[…]”, 14 October 2016, slide 42 [Doc ID 77].

NXP internal document, Email from […] titled “[…]”, dated 2 September 2015, [DOC ID 1458-51968].

NXP internal document, “[…]”, in particular slide 3-4 and 11-18; attachment to email from […] titled “[…]”, dated 2 September 2015, [DOC ID 1458-51968].

NXP internal document, email by […], dated 1 September 2015, “[…]”, [DOC ID 1458-16481].

(465) Therefore, through MIFARE4MOBILE, it is possible to develop transit services on mobile devices with MIFARE-enabled NFC/SE, which can interact with any transport infrastructures based on the various MIFARE standards, without the need for transit authorities to update their system or infrastructure to another MIFARE type. As explained in another NXP document, […] (emphasis added) and […] .

(466) As regards the development of mobile transit services, NXP’s internal documents highlight that, contrary to the Notifying Party’s views, the deployment of transit services through mobile devices is not merely speculative. For instance, the notes to an internal NXP presentation from 2016 highlight that […] (emphasis added) .

(467) NXP’s documents indicate that NXP has indeed taken steps to that end, and that MIFARE is crucial for the launch of mobile transit services. For instance, in a 2016 internal presentation discussing future goals, NXP identifies […] .

(468) In another document, prepared for a briefing on NXP’s products, NXP indicates that it has entered into a partnership with the Chinese OEM Xiaomi for mobile transit payments and that […]. As regards Europe, the document explains that […] .

(469) The same NXP document also explains that […] .

(470) In its internal documents, Qualcomm also discusses the importance of MIFARE, for device OEMs, NFC/SE providers and in relation to the development of transit services through mobile devices.

(471) For instance, in a Qualcomm internal email exchange on the Transaction’s synergies, MIFARE is qualified as a decisive factor for competing and as a […] as regards NFC and SE for mobile transit. In the discussion, Qualcomm’s […] explains that in relation to MIFARE, […]. Further highlighting the importance of MIFARE, the email comments that […] (emphasis added) .

(472) The email is also illustrative of the established presence of MIFARE, as it includes a snapshot, reproduced in Figure 4 below, that visualises the reach of NXP’s proprietary transit protocol.

Figure 4: MIFARE’s presence worldwide

[…]

Source: Qualcomm internal document, internal email exchange entitled […], email by […], dated 11 October 2016, [DOC ID 2476-882]

(473) Finally, the email also explains that the need for MIFARE is such to push device OEMs to purchase products that include it: […] .

(474) It also be noted that in an earlier draft email discussing MIFARE, Qualcomm’s […] explained that […]. This explanation is dismissive of the Notifying Party’s argument of the existence of credible alternative transit service technologies, as FeliCa is qualified as specific to Japan, as opposed to MIFARE (“global”), whereas other transit technologies are not even mentioned.

(475) Qualcomm internally acknowledged the importance of MIFARE also prior to the Transaction, when it entered into negotiations with NXP over a “eSE partnership” in 2013. When commenting on the topic of MIFARE, Qualcomm personnel noted that […]. In fact, in an earlier presentation, discussing the strategy for SE products, Qualcomm identified as a […] the need to […]. In another email exchange with a device OEM customer regarding SE chips, Qualcomm indicated that […] .

(476) More recently, in the context of the Transaction, in an email to Qualcomm’s […] emphasised the importance of NXP’s MIFARE technology for Qualcomm, commenting that […] (emphasis added). The presentation attached to the email also highlights the importance of MIFARE4MOBILE for Qualcomm’s strategy, as Qualcomm’s […] . Figure 5 below also illustrates that Qualcomm attached value to MIFARE4MOBILE and to its use for transit mobile services.

Figure 5: MIFARE4MOBILE architecture for mobile transit services

[…]

Source: Qualcomm internal document,[…] , October 2016, […]slide 11, [DOC ID 2360-76456] presentation attached to Qualcomm internal email by […], 31 October 2017 [DOC ID2364-69862]

(477) Market analysis reports also emphasise that MIFARE is the leading transit protocol technology. For instance, an ABI research on “Transportation Ticketing Technologies Market Update” from January 2015 states that […], despite a small decrease, and that […] . MIFARE’s importance is apparent from the current and forecasted high market share in terms of transit protocol competition, reproduced in Figure 6 below. The Commission also notes that in fact, MIFARE’s success is not

Qualcomm internal document, internal email exchange entitled “[…]”, email by […], dated 11 October 2016, [DOC ID 2476-882].

Qualcomm internal document, draft email by […] entitled “[…]”, email by […], dated 11 October 2016, [DOC ID 2364-7278].

Qualcomm internal document, internal email exchange entitled “[…]”, email by […] to […], dated 1 April 2013, [DOC ID 2476-2507].

Qualcomm internal document, “[…]”, slide 36, dated 11 February 2013 [DOC ID 2364-16303].

Qualcomm internal document, email by Qualcomm’s […] to […], entitled “[…]”, dated 16 May 2017, [DOC ID 2364-51626].

Qualcomm internal document, email by […] to […], 31 October 2017, [DOC ID 2364-69862].

Qualcomm internal document, “[…]”, October 2016, presentation by Qualcomm’s […] and […], slide 9, [DOC ID 2360-76456] presentation attached to Qualcomm internal email by […] to […], 31 October 2017 [DOC ID2364-69862].

NXP internal document, ABI research “Transportation Ticketing Technologies Market Update”, dated 2 January 2015, slide 3 [DOC ID 1334-6854].

linked to solely MIFARE Classic (which is actually the oldest type of MIFARE technology), but to various subsequent versions, such as DESFire and Ultralight.

Figure 6: MIFARE’s current and expected share in transit protocol competition

[…]

Source: NXP internal document, […], dated 2 January 2015, slides 1 and 3 [DOC ID 1334-6854]

(478) Another earlier ABI industry report also assessed that […] .

(479) The same report also comments that it very hard for new transit technologies to establish themselves, as entry is difficult and entry switching between transit protocols is uncommon: […]. In that regard, the ABI report also notes that alternative transit technologies such as FeliCa and Calypso have a more limited geographic footprint .

(480) In light of the above findings, the Commission therefore considers that NXP holds a dominant position in the market for transit service technologies through MIFARE, and that thus the merged entity would hold a dominant position within such market. MIFARE is the most relevant and distributed technology for transit services, in terms of installed base and shipments. Moreover, MIFARE is of great importance for device OEMs and NFC/SE providers for the purpose of mobile transit services, which are in the process of being deployed and being developed, among others, by the Parties themselves. Alternative transit service technologies, such as FeliCa and Calypso, do not seem to have the same presence and importance as MIFARE.

7.3.4. Market power with regard to IP (SEPs and non-SEPs)

7.3.4.1. Notifying Party's views

(481) The Notifying Party has indicated that it holds […] NFC patents (individual families or technologies) and […] granted and pending counterparts . Of these, the Notifying Party considers that it holds […] patent families that it believes to be potentially essential to the NFC standard.

(482) NXP holds owns approximately […] patent families, comprising […] individual patents and applications worldwide, that it believes to be relevant to NFC, SE, and SE OS. Of these, NXP believes five patent families (a total of […] individual U.S. patents, and […] individual EU patents) to be potentially essential to the NFC standard .

(483) In its reply to the Article 6.1(c) decision, the Notifying Party contends that, with regard to SEPs, NXP does not hold significant market power, as confirmed by the reply by one respondent to the phase I market investigation and by the fact that Samsung LSI, a competing supplier of NFC chips, does not have a license to NXP's NFC IP.

NXP internal document, ABI research report “contactless ticketing for transportation”, p. 6, 4 May 2012 [DOC ID 1334-4022].

NXP internal document, ABI research report “contactless ticketing for transportation”, p. 10, 4 May 2012 [DOC ID 1334-4022].

NXP internal document, ABI research report “contactless ticketing for transportation”, pp. 15-17, 4 May 2012 [DOC ID 1334-4022].

(484) As to non-SEPs, the Notifying Party notes that the IP portfolios of the Parties are complementary and do not give rise to horizontal concerns.

(485) In any event, the Notifying Party notes that the Parties have not carried out an estimate of their shares of the declared potential NFC SEPs or NFC SEPs and that the exact number of patents is not a meaningful basis for the assessment of market power .

7.3.4.2. The results of the market investigation and the Commission's assessment

(486) In line with its decisional practice mentioned in Section 5.2.5 above, the Commission considers that – in relation to NFC technology – each of NXP and Qualcomm hold market power with regard to their NFC SEPs since standard implementers in principle cannot design around them and thus "a potential licensee […] cannot switch to another supplier" . The phase I and phase II market investigation has not revealed any reason for departing from the previous Commission's case practice .

(487) As to the Notifying Party's arguments, the Commission notes that Qualcomm mischaracterized the statement of the respondent to the phase I market investigation. First, the same respondent considered that NXP's NFC SEPs are necessary to comply with the NFC standard and cannot be designed around. Second, it indicated that the industry lacks a specific interest in NXP's NFC SEPs not because (as the Notifying Party claims) of their lack of importance, but because NXP does not require its customers buying NFC chips to take a patent license (described as "selling exhaustively"), which therefore has not solicited an interest to investigate NXP's patent portfolio .

(488) As to Samsung's absence of license to NXP's IP portfolio, […] .

(489) As to NFC non-SEPs, a small number of respondents to the phase II market investigation provided an answer with regard to the importance of NFC non-SEPs, indicating that they are either important or very important . However, the Commission considers that, for the purpose of the present Decision, it is not necessary to reach a conclusion as to whether NXP holds market power with regard to its NFC non-SEPs.

(490) The Commission therefore concludes that post-Transaction the merged entity would hold a significant degree of market power as regards the NFC SEPs of Qualcomm and NXP.

7.4. Conglomerate effects in relation to Qualcomm's baseband chipsets and NXP's NFC and SE chips

7.4.1. Introduction

(491) In Section 7.4, the Commission will analyse whether the Transaction would give rise to conglomerate concerns in relation to Qualcomm's LTE baseband chipsetsand NXP's NFC chips, SE chips, NFC/SE combined solutions and MIFARE.

(492) As explained in Section 7.1, the Non-Horizontal Merger Guidelines distinguish between bundling, which usually refers to the way products are offered and priced by the merged entityand tying, usually referring to situations where customers that purchase one good (the tying good) are required to also purchase another good from the producer (the tied good). While tying and bundling have often no anticompetitive consequences, in certain circumstances such practices may lead to a reduction in actual or potential competitors' ability or incentive to compete. This may reduce the competitive pressure on the merged entity allowing it to increase prices .

(493) When conducting its analysis, the Commission examines first, whether the merged firm would have the ability to engage in a given conduct, second, whether it would have the economic incentive to do soand, third, whether that conduct would have a significant detrimental effect on competition, thus causing harm to consumers. In practice, these factors are often examined together as they are closely intertwined.

(494) In the Article 6(1)(c) decision, the Commission raised serious doubts as to the compatibility of the Transaction with the internal market on the grounds that the merged entity would have the ability and incentive to engage in technical and contractual tying, pure and mixed bundling and degradation of interoperability in relation to baseband chipsets, NFC and SE chips, which would have likely led to the foreclosure of standalone providers of those products.

(495) Following the phase II market investigation, the Commission considers that the merged entity would be likely to engage in mixed bundling with regard to Qualcomm's LTE baseband chipsets and NXP’s NFC and (MIFARE-enabled) SE chips. The Commission however does not consider it likely that this conduct in itself would lead to significant anti-competitive effects.

(496) However, the Commission considers it likely that, in addition to engaging in a foreclosure strategy of mixed bundling, the merged entity would also have the ability and incentive to raise the licensing royalties for MIFARE or, in the extreme, to cease the licensing of MIFARE. This conduct, coupled with mixed bundling, would lead to foreclosure of third party providers of baseband chipsets and NFC and SE products.

(497) Following the phase II market investigation, the Commission also considers that the merged entity would have the ability and incentive to engage in a conduct of degradation of interoperability of third parties’ baseband chipsets and NFC and SE products. This conduct would compound the foreclosure effects stemming from mixed bundling and increase in MIFARE royalties.

(498) Finally, following its phase II market investigation, the Commission considers that the merged entity would be unlikely to engage in conducts consisting of pure bundling or technical and contractual tying of its baseband, NFC and SE products or both. Furthermore, in the event that the merged entity were to engage in such conducts, the Commission considers that these conducts would be unlikely to lead to significant foreclosure effects.

(499) The Commission will first carry out its conglomerate assessment in relation to a potential mixed bundling conduct, including in relation to the licensing of MIFARE (Section 7.4.2), then in relation to potential pure bundling and tying conducts (Section 7.4.3), and finally in relation to a conduct of possible degradation of interoperability (Section 7.4.4).

7.4.2. Mixed bundling and raising MIFARE royalties

7.4.2.1. Background

(500) In Section 7.4.2, the Commission will assess conglomerate effects in relation to a mixed bundling conduct concerning NXP's NFC and SE products, including MIFARE, and Qualcomm’s LTE baseband chipsets, and will consider what the likely impact on competition of such conduct would be.

(501) Bundling can take two forms, pure bundling and mixed bundling. In the case of pure bundling, the products are only sold jointly in fixed proportions. Under mixed bundling, the products are also available separately, but the sum of the stand-alone prices is higher than the bundled price .

(502) Therefore, for the assessment of the present case, in a mixed bundling scenario the merged entity would offer a bundle comprising those products at a discount compared to the sum of the prices of the standalone products. The merged entity would continue to offer the standalone products alongside the bundle, at a higher price than the bundle.

(503) At the outset, the Commission recalls that, as indicated in the Non-Horizontal Merger Guidelines, in order to be able to foreclose competitors, the new entity must have a significant degree of market power, which does not necessarily amount to dominance, in one of the markets concerned. The effects of (mixed) bundling can be expected to be substantial when at least one of the merging parties' products is viewed by many customers as particularly important and there are few relevant alternatives for that product .

(504) In this respect, with regard to market power of the merged entity, the Commission considers that the evidence on the file as presented in Section 7.3 shows that post-Transaction the merged entity would indeed hold a dominant position in the worldwide market for LTE baseband chipsets and in the market for transit service technologies, and a certain degree of market power in the markets for NFC, SE and NFC/SE combined solutions.

(505) With regard to the dominance of the merged entity within LTE baseband chipsets, reference is made to the arguments and findings of Section 7.3.1.2 in its entirety.

See Non-Horizontal Merger Guidelines, paragraph 96.

See Non-Horizontal Merger Guidelines, paragraph 99.

(506) With regard to market power of the merged entity with respect to NFC chips, SE chips, and NFC/SE combined solutions reference is made to the arguments and findings of Section 7.3.2.2 in its entirety.

(507) With regard to the dominance of the merged entity with respect to transit service technologies, reference is made to the arguments and findings of Section 7.3.3.2 in its entirety.

(508) Based on the aforementioned findings, the Commission considers that the merged entity’s products, in particular LTE baseband chipsets but also NFC/SE chips and MIFARE, would be important products for device OEM customers.

(509) Further, as stated in the Non-Horizontal Merger Guidelines, for a conduct that results in foreclosure to be of potential concern, it must be the case that there is a large common pool of customers for the individual products concerned. The more customers that tend to buy both products, the more demand for the individual products may be affected through bundling. Such a correspondence in purchasing behaviour is more likely to be significant when the products in question are complementary (emphasis added) .

(510) However, the Non-Horizontal Merger Guidelines also indicate that customers may have a strong incentive to buy the range of products concerned from a single source ("one-stop-shopping") rather than from many suppliers, for example because it saves on the transaction costs. The fact that post-Transaction the merged entity would have a broad range or portfolio of products does not, as such, raise competition concerns .

(511) As pointed out by the Non-Horizontal Merger Guidelines, the merged entity's incentive to foreclose rivals through bundling would depend on the degree to which such strategy is profitable for the merged entity .

(512) The Commission will assess mixed bundling against this backdrop.

7.4.2. Mixed bundling and raising MIFARE royalties

7.4.2.1. Background

(513) Based on the results of the phase II market investigation, the Commission considers that post-Transaction the merged entity would have the ability and incentive to engage in mixed bundling of its LTE baseband chipsets with NXP’s NFC and SE products, but that this strategy in itself is unlikely to have significant anti-competitive effects.

(514) Yet, the Commission also considers that post-Transaction the merged entity would have the ability and incentive to add MIFARE to the aforementioned mixed bundle and couple such mixed bundling with an increase in the royalty rate of MIFARE, or to refuse to license MIFARE altogether, and that this strategy would be likely to have significant anti-competitive effects.

Non-Horizontal Merger Guidelines, paragraph 100.

Non-Horizontal Merger Guidelines, paragraph 104.

(515) As a first step of this conduct, the merged entity would have the ability and incentive to engage in a mixed bundling strategy, thereby offering the Parties’ products commercially bundled together at a discount compared to the sum of the prices of those stand-alone components. The bundle would consist of the Qualcomm LTE baseband chipset, together with the NXP NFC/SE combined solution, enabled with MIFARE, and the price of such bundle would be lower than the sum of the prices of the respective standalone components.

(516) As a second step in its mixed bundling strategy, the merged entity would have the ability and incentive to technically integrate NXP’s MIFARE-enabled SE on the LTE baseband chipset (the Snapdragon platform). Following such integration, Qualcomm would offer to device OEMs both a bundled product comprising the LTE baseband chipset (integrated with MIFARE enabled SE) and the NFC controller, and a set of standalone components, where the bundled product would sell at a discount compared to the sum of the prices of the stand-alone components.

(517) In parallel and in addition to the steps described above, the merged entity would have the ability and incentive to degrade the conditions of access to MIFARE for other NFC/SE suppliers, by raising the licensing royalties or in the extreme by ceasing the licensing of MIFARE altogether. Given the importance and relevance of MIFARE as a technology protocol for mobile transit services, this conduct would result in significant anticompetitive effects against the backdrop of a mixed bundling strategy. As will be discussed in recitals (668) et seq. below, should Qualcomm enable MIFARE on its own SEs and license it under terms which make it unprofitable to other competitors (or deny it completely), such conduct would increase the risk of foreclosure of both rival baseband and NFC/SE suppliers, as their products or bundles (without MIFARE or with an increased MIFARE royalty) would be less competitive against the merged entity and less attractive to device OEMs.

(518) Strong concerns that Qualcomm would pursue such bundling strategy coupled with a degrading of conditions to license MIFARE were raised during the market investigation, in particular by manufacturers of NFC/SE chips and by customer OEMs.

(519) The aforementioned conduct would likely result in decreased profitability and reduced market shares of the merged entity’s competitors, leading to foreclosure effects, as standalone competitors’ ability and incentive to compete would be reduced. Given that both the baseband chipset and NFC and SE markets are characterised by intensive R&D activity and investment in R&D, declining market shares and profits may therefore reduce prospects to grow and monetize innovations, which would discourage both its baseband and NFC and SE competitors to continue investing in developing the products.

(520) This conduct is discussed in more detail in what follows.

A) Notifying Party’s views

(521) At the outset, the Notifying Party recalls that mixed bundling in itself is presumptively pro-competitive.

(522) As regards the ability to engage in a foreclosure strategy of mixed bundling, the Notifying Party argues the following.

(523) First, the Notifying Party submits that the merged entity would not hold market power in relation to LTE baseband chipsets and NFC and SE chips or MIFARE and that it would not have the ability to foreclose rival suppliers by engaging in a strategy of mixed bundling of sales of its LTE baseband chipsets with NXP's NFC or SE chips or both. As regards baseband chipsets, the Notifying Party faces competition from several manufacturers, including vertically integrated device OEMs such as Samsung and Huawei. With respect to NFC and SE, the merged entity would face competition from other providers of NFC and SE, as well as from providers of alternative technologies to NFC (such as BLE) and SE (such as HCE) post-Transaction. The Notifying Party also considers that mobile device OEMs can and do source baseband chipsets and NFC and SE chips from a variety of other suppliers and that the merged entity would be unable to foreclose rivals through bundling.

(524) Second, the Notifying Party considers that there are practical impediments to practices such as bundling: LTE baseband chipsets and NFC and SE chips are purchased through different procurement teams, and at different times. At a minimum, the fact that customers do not currently structure their tendering processes in a manner compatible with LTE baseband chipset and NFC/SE bundling poses an obstacle to the use of a bundling strategy by the merged entity and serves to make such a strategy less viable .

(525) Third, the Notifying Party notes that in analogous circumstances where it has acquired companies producing complementary chips, there is no evidence that a conglomerate strategy of bundling or combining products led to foreclosure. For instance, despite offering Wi-Fi and Bluetooth chips in combination with its baseband chipsets, there is no evidence that the Notifying Party has been able to dominate these markets or to foreclose competitors (the Notifying Party holds only [20-30]% and [20-30]% shares of Bluetooth and Wi-Fi supply respectively) .

(526) Fourth, in relation to MIFARE, the Notifying Party claims that MIFARE is not a "must have" product and that there is virtually no implementation of MIFARE for transit systems in mobile phones. Then, there are available alternatives to MIFARE, such as EMV or QR codes. Further, various rival suppliers of NXP have ongoing licenses for MIFARE, which are due to expire at the earliest in […]. This means that those suppliers are protected for a few more years from any conduct by the merged entity.

(527) The Notifying Party also submits that it does not have the incentive to engage in a mixed bundling strategy. Its incentive to engage in mixed bundling and to increase the price of standalone products is limited by the specific facts of the markets for LTE baseband chipsets and NFC/SE.

(528) First, the fact that NFC/SE and baseband chipsets account for a small proportion of handset input costs limits the scope of the "Cournot effect" between the two, which in turn would limit the incentive for the merged entity to reduce the price of the LTE baseband chipsets to drive increased usage of NFC/SE and vice versa.

(529) Thus, Cournot effects are unlikely to be material in respect of BC and NFC/SE supply. The extent to which a merged firm would be willing to reduce the price of component products (for example, BC) as a result of this effect is driven by two factors. First, the sensitivity of demand for BC to a change in the price of BC relative to the sensitivity of demand for BC (NFC/SE) to a change in the price of NFC/SE (BC). Second, the relative BC and NFC prices.

(530) In relation to the first point, BC and NFC/SE prices represent only a relatively small percentage of the total price of a mobile phone, and the price of NFC chips is even lower. This implies that it is unlikely that a 5-10% reduction in the price of BC (NFC) would result in a significant increase in mobile phone sales and a resulting expansion in NFC/SE (BC) sales. As such, it appears unlikely that the sensitivity of demand for the BC to a change in the price of the associated NFC will be significant (relative to its own price elasticity) and vice versa.

(531) Second, price negotiation is feasible without engaging in bundling: the use of detailed, individualised price negotiations with OEM customers means that the Parties would not be constrained by the need to offer a uniform price to all customers. This would reduce the importance of implicit price discrimination as a potential motivation for mixed bundling post-Transaction.

(532) Finally, the Notifying Party argues that even if it would have the ability and incentive to engage in such bundling practices post-Transaction , this would have no or only negligible effects on competition.

(533) First, the Notifying Party argues that the market is characterised by countervailing buyer power by large mobile device OEMs: the merged entity's customers are large, well informed OEMs, which are able to take a strategic view of the market. These customers will likely not accept a bundle that could be seen as lessening future competition in the NFC/SE and LTE baseband chipset markets. Furthermore, those OEMs may value purchasing standalone components from rival suppliers (following a "mix-and-match" approach) and would not see either of the Parties' components as unique or "must have", so that they would not choose the bundle.

(534) In that regard, large mobile device OEMs would find it rational to act strategically to protect competition, particularly where they have an existing relationship with a second supplier or in-house manufacturing capabilities. OEMs have sufficient strategic business insights and negotiating power to anticipate any future market trends, and continue purchasing the most cutting-edge technologies at the lowest prices.

(535) Second, the Notifying Party submits that, even if its current customers decided to purchase the products of the merged entity, there would still be a very large amount of new NFC demand: as NFC/SE products become cheaper and more pervasive, demand for such products will extend to those OEMs that do not purchase NFC/SE technology today. This expansion in demand would provide additional sales opportunities to the standalone competitors of the merged entity, which could address such demand, thus justifying their continued investments and presence in NCF/SE technology in the future. In fact, according to the Notifying Party, the anticipated growth in NFC demand will provide sufficient sales opportunities for those products, unrelated to Qualcomm’s BC supply, which will be sufficient to sustain rival NFC/SE suppliers and therefore effective competition.

(536) More specifically, the Notifying Party submits that its baseband chipsets are incorporated in [60-70]% of mobile phones which contain an NFC chip, but in only [20-30]% of mobile phones without an NFC chip. Volumes of mobile phones without NFC chips represents the potential demand growth for NFC chips once NFC becomes a standard feature on mobile phones of all tiers. For 2016, handset devices without NFC totalled […], of which […] did not incorporate a Qualcomm baseband chip, and would therefore not be affected by any bundling strategy of Qualcomm baseband chipsets with NFC chips after the Transaction.

(537) Furthermore, the Notifying Party submits that specific NFC/SE suppliers would be protected, limiting any anti-competitive effects brought by the Transaction.

(538) Samsung LSI is the second largest current supplier of NFC chips and supplies predominantly its own downstream arm, Samsung Mobile. Samsung would thus find it irrational to accept a bundle offered by the merged entity that would undermine its own business going forward. Therefore it can be expected that Samsung LSI would remain a competitor to Qualcomm, supported by ongoing sales to Samsung Mobile. It is unrealistic to expect that any mixed bundling discount would offset the benefit that Samsung obtains from purchasing NFC from its own upstream arm. Sony Semiconductors is the third largest current supplier of NFC and currently sells NFC to its downstream arm only. For similar reasons to those outlined for Samsung, the Notifying Party considers it implausible that NXP could encourage Sony to undermine its own in-house supply. The Notifying Party considers that the same argument holds for HiSilicon, a subsidiary of the device OEM Huawei, which currently produces and sells SE chips to Huawei.

(539) STMicroelectronics is becoming an increasingly important competitor having invested significantly in research and development following its acquisition of AMS in 2016. STMicroelectronics already supplies the SE that is used in the Apple SIM. This relationship with Apple is further likely to protect its competitive position in the future.

(540) Finally, the Notifying Party also considers that the issue that it would leverage MIFARE to exclude rival SE chips manufacturers is not specific to this merger, as both MIFARE and SE are existing NXP products, and NXP could already engage in a bundling and royalty-increasing conduct today.

B) Commission's assessment

1. Ability to engage in mixed bundling

(541) Following its phase II market investigation, the Commission considers that post-Transaction the merged entity would have the ability to engage in a mixed bundling strategy of Qualcomm’s LTE baseband chipsets with NXP’s NFC/SE combined solution, and MIFARE. In addition, Qualcomm would have the ability to raise the licensing royalties for MIFARE or cease to license MIFARE altogether to customers and competitors.

(542) More specifically, the Commission considers that, as a first step in its strategy, the merged entity would be clearly able to offer a commercial bundle consisting of its LTE baseband chipset and NXP's NFC/SE combined solution (with MIFARE) at a discount price compared to standalone LTE baseband chipsets, and standalone NFC/SE combined solutions (with MIFARE). The Commission considers that this conduct could be implemented immediately after the Transaction, as the merged entity would not face technological hurdles when offering the products commercially together.

(543) The merged entity would offer those products in this formula in the first few years after the Transaction, which would correspond to the period needed to integrate the NXP SE on the Qualcomm baseband chipset.

(544) As a second step, within the next two to three years after the Transaction, the merged entity would have the ability to technically integrate NXP’s SE with the Qualcomm LTE baseband, and would have the ability to offer as a discounted bundle the integrated LTE baseband-SE product (implementing MIFARE) with the NFC chip, at the same time raising the price of the standalone integrated products. In parallel, Qualcomm would continue having the ability to ask for higher royalties for MIFARE in licensing negotiations or of no licensing altogether.

(545) Based on the findings of the phase II market investigation, the Commission considers that the merged entity would have the ability to engage in the aforementioned conduct for the following reasons.

(546) First, LTE baseband chipsets and NFC/SE chips and MIFARE are complementary products and are purchased by a common pool of customers which are the mobile device OEMs .

(547) Based on the information provided by the Notifying Party, the Parties have a common pool of customers: nine out of Qualcomm's top ten baseband chipset customers, as well as Apple, also purchase NFC and SE chips from NXP or other suppliers; similarly, NXP's top ten NFC device OEM customers also source baseband chips from Qualcomm or other suppliers. These customers can hence be addressed by a mixed bundling strategy.

(548) The Commission considers that Qualcomm's baseband chipsets and NXP's NFC/SE products can be considered complementary within the meaning of paragraph 91 of the Non-Horizontal Merger Guidelines. Device OEM customers buy these products as inputs for their mobile devices.

(549) The Commission notes that the merged entity would face no difficulty in promptly implementing the first step of the mixed bundling conduct after the Transaction, given that it would not face any technological hurdles to do so, since the commercial variant of mixed bundling does not require any physical modification of the products included in the bundle. To the contrary, the fact that LTE baseband chipsets and NFC and SE chips, as well as MIFARE, are complementary products that are purchased by a common pool of customers facilitates the adoption of a mixed bundling strategy. MIFARE could also be easily added to the bundle.

(550) Second, respondents to the Commission’s phase I market investigation indicated that Qualcomm would have the ability to engage in mixed bundling of baseband and NFC/SE products. Those respondents also indicated that Qualcomm would have the means and ability to technically integrate NXP’s products (in particular, the SE chip) on the Qualcomm baseband, and to refuse licensing MIFARE (or license MIFARE at increased royalties).

See Non-Horizontal Merger Guidelines, paragraph 100.

According to Qualcomm, Apple is an indirect Qualcomm baseband chipset customer in that its devices are made by contract manufacturers, which source chipsets from suppliers like Qualcomm and pass-on their costs to Apple.

Form CO, paragraph 918, Tables 4.4 and 4.5 [DOC ID 326].

Form CO, paragraph 919, table 4.6 [DOC ID 326].

See responses to question 65 of Q1 – Questionnaire to device OEMs; responses to question 56 of Q2 – Questionnaire to baseband competitors; responses to question 62 of Q3 – Questionnaire to NFC competitors.

See response by […] to question 65 of Q1 – Questionnaire to device OEMs, [DOC ID: 1009].

See response by Microsoft to question 65 of Q1 – Questionnaire to device OEMs. [DOC ID: 776]

See for instance responses to question 3.2 of RFI 29 [DOC ID: 2955].

See response by Samsung to question 19 of RFI 28 [DOC ID 2671].

See response by Samsung to question 7 of the Commission’s supplemental questions to RFI 28 [DOC ID: 2672].

Third party submission of 2 October 2017, page 9 [DOC ID 3138].

110

with Qualcomm with respect to the solution as a whole, and does not negotiate on a chip-by-chip basis .

(556) Therefore, it appears that Qualcomm could implement a mixed bundling strategy of NFC/SE chips around the sale of the LTE baseband chipset.

(557) With regard to the second step in the mixed bundling strategy, namely the ability to technically integrate the SE chip on the LTE baseband, respondents to the phase I market investigation, both among customers and competitors, indicated that Qualcomm would have the technical ability to engage in product integration of the SE chip with the baseband chipset.

(558) The Commission recalls that, during the phase I investigation, respondents among customers and competitors indicated that in their view the merged entity would have the ability to engage in various types of product integration, notably not just the SE chip with the Qualcomm baseband, but also the NFC chip with the baseband chipset .

(559) Following the phase II market investigation, the Commission considers that, while it may be possible for the merged entity to also integrate the NFC chip with the Qualcomm baseband, this conduct does not appear to be one that the merged entity would most likely pursue in the near future. Rather, the most likely and feasible type of technical integration that the merged entity would be likely to pursue in the near future would be the integration of the SE chip with the Qualcomm baseband, as this conduct is the one that, in the view of respondents to the market investigation illustrated in the following recitals, is the most likely and achievable.

(560) Moreover, as is also evidenced by the Parties’ internal documents further discussed below in this Section, the type of integration that the Parties mainly discussed and would consider implementing would be that of the SE chip on the Qualcomm baseband.

(561) Some respondents, such as Infineon, indicated that Qualcomm has already presented at the 2017 GSMA Congress in Barcelona a prototype integrating a baseband chipset and a secure element. Although this prototype is focused on the integration of the secure iUICC within the Snapdragon baseband chipset, this shows that Qualcomm does pursue a similar integration strategy and would have the ability to integrate the SE on the Snapdragon.

(562) Gemalto submitted that Qualcomm is most likely to “(a) cease licensing MIFARE and make MIFARE enabled SEs available bundled with their own basebands and /or AP chipsets; or (b) move to an approach where SEs are integrated with basebands and/or APs and where Qualcomm no longer offers MIFARE-enabled NXP SEs on a standalone basis” .

Samsung’s reply to the Commission’s RFI 28, question 19, [DOC ID 2671].

Integration of the NFC chip in the baseband chipset, integration of the SE in the application processor, integration of the SE in the baseband chipset. See Responses to question 52 of Q1 – Questionnaire to device OEMs; Responses to question 45 of Q2 – Questionnaire to baseband chipsets competitors and Responses to question 52 of Q3 – Questionnaire to NFC competitors.

NXP internal document, “[…]”, email sent by Qualcomm’s […], dated 9 May 2017, [DOC ID 1458-30755] [[…] msg].

Qualcomm internal document, email by […], entitled “[…]”, dated 4 November 2016, [DOC ID 2364-72772].

NXP internal document, “[…]”, word document attached to email from […] to […], “[…]”, 3 May 2017, [DOC ID 1456-52124] […].

NXP internal document, email from […] to […], “[…]”, dated 26 May 2017, [DOC ID 1456-53926] […].

111

(563) Among the customers, […] .

(564) Samsung also submitted that it may be possible for Qualcomm to technically integrate baseband chipsets and secure element, but could not speculate on the costs and time involved for such process .

(565) Respondents to the phase I market investigation also pointed out that the merged entity would have the ability to bundle a MIFARE-enabled SE with Qualcomm's BC and offer them to customers that would otherwise buy these products separately . One customer OEM indicated that, in a situation where the merged entity would bundle these products, customers would not be able to turn to alternative technologies such as FeliCa or Calypso, as MIFARE is the dominant technology in contactless ticketing worldwide. The phase II market investigation confirms this view .

(566) Third, this mixed bundling approach of first offering the Parties’ products (baseband, NFC/ (MIFARE enabled) SE) commercially together and subsequently offering them in a technically integrated mixed bundle (in particular, the SE on the Qualcomm baseband) is also reflected in the Parties’ internal documents.

(567) As regards the commercial offering of the Parties’ products together at a discount, an internal document of the Parties, related to the “go to market approach” of the merged entity, emphasises as key considerations to […]. The same internal document also indicates that […] .

(568) Indeed, it appears that in the course of several integration summits, the Parties outlined plans to ensure […] . The internal documents prepared for those integration summits illustrate that the Parties aimed to ensure the feasibility of cross-selling of NXP’s products to Qualcomm device OEM customers, in particular NFC products, as of day one after the Transaction as can be seen in Figure 7 below.

Figure 7: Parties’ checklist of feasibility of cross-selling of NXP NFC products

[…]

Source: Qualcomm internal document, […] .

(569) Moreover, the Parties’ internal documents also show that NXP’s sales representatives for products to device OEMs were expected to be moved to Qualcomm’s mobile sales teams dealing with the device OEMs. In an email exchange between the Parties’ representatives and the external consultants of […], discussing the post-merger organisation of sales, it can be read that “[…]” (emphasis added) . Therefore, under this sales structure (agreed between the Parties), contrary to the Notifying Party’s argument mentioned in recital (554) above, it would not be

(580) Finally, as MIFARE is a NXP proprietary technology, the merged entity would have the ability to raise royalties or cease the licensing of MIFARE once the ongoing licenses expire. NXP has not taken any obligations with any standard setting organisation to license this technology on F(RAND) terms or to license it at all . Outside the obligations taken as part of the ongoing licensing agreements, NXP, and post-Transaction, the merged entity, does not have an obligation to license MIFARE to other suppliers of NFC/SE solutions after the expiry of the ongoing licensing agreements. At the expiry date of these agreements, the merged entity could be the only provider of the MIFARE technology if it chooses so. Therefore, as of […], the merged entity (when current MIFARE agreements have ended) would have the ability to stop licensing MIFARE, once its current contractual obligations come to an end.

(581) In light of the above, the Commission therefore takes the view that the merged entity would have the ability to engage in a mixed bundling strategy of Qualcomm's LTE baseband chipsets with NXP's NFC and (MIFARE-enabled) SE chips and to degrade licensing conditions for MIFARE (or cease licensing MIFARE altogether). It would also have the ability to technically integrate at least the SE chip on its baseband chipset in the foreseeable future.

The Linley Group: "Mobile Semiconductor Market Share Forecast 2015-2020", page 50 and 53; Form CO, Annex 4.16 [DOC ID: 327].

The Linley Group: "Mobile Semiconductor Market Share Forecast 2015-2020", page 61; Form CO, Annex 4.16 [DOC ID: 327].

Samsung's reply to RFI 28 of 22 July 2017, question 19, [DOC ID: 2671].

Qualcomm internal document, internal email exchange entitled “[…]”, email by […], dated 11 October 2016, [DOC ID 2476-882].

See the Notifying Party's response to Commission's RFI of 28 February 2017, question 22.

Gemalto, ST Microelectronics or G&D have ongoing MIFARE licensing agreements with NXP, these agreements will come to an end within […]; Gemalto and G&D have licensing agreements which are due to expire in […], and STMicroelectronics has an agreement which expires in […].

(582) For the reasons set out in the following recitals, the Commission considers that post-Transaction the merged entity would have the incentive to engage in mixed bundling and to increase MIFARE royalties (or in the extreme refusal to license MIFARE), in the fashion outlined in recitals (513) and (517).

(583) With regard to the first step of the mixed bundling strategy, concerning the sale of LTE baseband chipsets together with NFC and (MIFARE-enabled) SE chips at a discount compared to sum of the prices of the relevant stand-alone components, the Commission considers that the merged entity would have the incentive to engage in such conduct for the following reasons.

(584) First, as explained in recitals (546) to (549) above, the Commission considers that Qualcomm's baseband chipsets and NXP's NFC/SE products can be considered complementary within the meaning of paragraph 91 of the Non-Horizontal Merger Guidelines: mobile device OEM customers tend to buy all those products for their end devices, with baseband chipsets clearly being indispensable for the functioning of mobile devices, and NFC/SE chips and MIFARE increasingly becoming more and more important for such devices.

(585) Second, the Commission notes that mixed bundling of complementary products would very likely be a profitable strategy for the merged entity even in the short run, irrespective of whether foreclosure occurs. The merged entity would have an incentive to engage in mixed bundling, setting the price of the bundle composed of its own products below the sum of the two standalone prices. This is because the merged entity now internalises the positive externality that a lower price of one component generates in terms of demand for other complementary components. At the same time, post-Transaction the merged entity would also have an incentive to raise the price of its standalone components (that is, BCs or NFC/SE chips), thus making its competitors' products less attractive to those customers who want to "mix and match", that is to buy one component from the merged entity and the complementary component from a third-party producer. This price increase of the standalone components would divert demand away from such mix-and-match solutions towards the bundle offered by the merged entity. Optimising the pricing of the own bundle and of the stand-alone components in this manner could be expected to increase the sales of the merged entity, making mixed bundling a profitable strategy .

(586) Third, when asked whether Qualcomm would have the incentive to engage in mixed bundling by offering the bundle of baseband chipsets and NFC/SE products (including MIFARE) at lower prices than the standalone products, a majority of respondents to the phase I market investigation confirm that Qualcomm indeed has that incentive . […]. Gemalto considers that in the event of mixed bundling Qualcomm could, rather than lowering the price of the bundle "increase the price of the standalone products to expand its demand for the bundle". Samsung explains that "[a]ssuming that such bundled price is indeed cheaper than the prices of standalone products available from competitors, then OEM[s] most likely would buy the cheaper Qualcomm bundled product. However, such a price likely will not persist, if it occurs at all. The likely effect would be that Qualcomm would drive out competition from standalone product suppliers and then later raise the price after competitors are driven out of the relevant markets" .

(587) Furthermore, most respondents to the phase I market investigation who expressed an opinion also considered that in a situation of mixed bundling, Qualcomm would have the incentive to charge higher prices for stand-alone products compared to pre-Transaction prices by Qualcomm and NXP respectively . For instance, one respondent to the phase I market investigation submitted that the merged entity would […] "substantially raise the price of the NXP chips, and offer a very significant discount to customers who also buy Qualcomm chips; and […] offer a very significant discount on Qualcomm chips to customers who also buy NXP chips". Gemalto explains: "[i]t is the possibility of bundling the two products that makes it profitable post-merger to increase the price of the separate products compared to the prices in the absence of the merger. This would increase demand for the bundle and margins on the standalone products" .

(588) Some customers explained that they could not ignore the advantages of the bundled products. Samsung explained that to a certain degree it already uses some competing chipsets (baseband included) to the extent it this does not result in too large disadvantages. However, "Samsung is continuously under pressure from network carriers, who are Samsung’s direct customers, to lower the materials costs and the price of Samsung’s handsets – even at the high-end and flagship tiers. Therefore, Samsung has minimal flexibility in how much it can divert business to other suppliers – assuming that it has that choice in a particular market – especially if the merged company would offer “a bundle at a cheaper price" .

(589) Fourth, as mentioned in recitals (528) to (530), the Notifying Party argues that the fact that NFC/SE and baseband chipsets account for a small proportion of handset input costs limits the scope for discounts on the Parties’ own bundle in a mixed bundling strategy after the Transaction. The Notifying Party claims that any such discount, even if sizable as a fraction of the value of the bundled products, would only lead to a small reduction (if any) in the final price of the mobile device containing these bundled products, implying that final consumers are unlikely to switch mobile device brands as a result of such small changes in mobile device prices. The Notifying Party therefore concludes that mixed bundling would have a very small (or even no) impact on demand for their own bundle (or demand for the third parties' components), so that they would not have any incentive to engage in that pricing strategy in the first place.

(590) This claim is based on the mistaken premise that the impact of mixed bundling on the demand for the Parties' products operates through final consumers, and through their choice of mobile device brands. It may indeed be the case that consumers' demand for any given mobile device brand is insensitive to the wholesale price of mobile device components. However, the demand for LTE baseband chips and NFC/SE solutions that is relevant to the merged entity is the wholesale demand of device OEMs, who are the direct customers of the component producers, and not the retail demand of final consumers of the mobile devices. As pointed out by the Notifying Party, device OEMs are likely to carefully evaluate the procurement cost of any given mobile device component before placing their orders. Given the significant level of expenditure on LTE baseband chipsets and NFC/SE solutions in the total input cost of device OEMs, their demand for any given product is likely to be sensitive to its prices. It is therefore ingenious to claim that a mixed bundling strategy would have no impact on the demand for the Parties' products, and could therefore not be profitably pursued.

(591) Fifth, the Parties’ internal documents suggest that the merged entity would have an incentive to engage in mixed bundling of LTE baseband chipsets with NFC and (MIFARE enabled) SE chips, and in the integration of the SE chip on the baseband chipset.

(592) In documents discussing the Transaction's synergies, the Notifying Party mentions the […] to Qualcomm and […], including […]. Another Qualcomm internal document discussing the […] of the Transaction mentions the […] .

(593) Other internal documents discussing the “Go To Market” strategy of the merged entity illustrate the […], which includes among the mentioned product categories NXP’s SMT (“Secure Mobile Transactions”) line of business, which encompasses NXP’s NFC and SE chips, as can be seen in Figure 8.

Figure 8: Parties’ estimated revenues from cross-selling to mobile customers

[…]

Source: NXP internal document,[…] .

(594) Similarly, in another internal document the Parties clearly indicate among the […] the action to […] and to […]. Another internal document also indicates that Qualcomm’s approach with regard to the mobile space is indeed to engage in cross selling, as reproduced in Figure 9. A subsequent slide of the same document acknowledges that there are […] .

Qualcomm internal document, "[…]", 19 December 2016, slide 4, Annex 9.1 to the Form CO.

Qualcomm internal document, "[…]", 1 May 2016, slide 10, Annex .4.b.c.5 to the Form CO.

Presentation attached to Email sent by NXP’s […] internally within NXP [Doc ID 1456-28230].

NXP internal document, “[…]”, 19 January 2017, slide 4, presentation attached to Email sent by NXP’s […] internally within NXP [Doc ID 1456-28230] [[…].msg].

Source: […], slide 10, dated 7 February 2017.

Figure 9: revenues from mobile cross-selling

[…]

Source: […]

(595) The internal documents of the Parties also provide strong indications that Qualcomm would have the incentive to leverage MIFARE in order to protect the core business of baseband chipsets. MIFARE is also the product which would give the merged entity an advantage over other NFC/SE suppliers. In an email entitled […] addressed to other executives in Qualcomm, Qualcomm's […] wrote: […]. To which Qualcomm's […] replied: […](emphasis added).

(596) In another correspondence between NXP and Qualcomm, titled […], NXP pointed out that they also considered adding MIFARE as a potential cross sell with Qualcomm's mobile platform .

(597) In light of the findings in recitals (583) to (596), the Commission therefore considers that the merged entity would have the incentive to engage in mixed bundling of LTE baseband chipsets with NFC and (MIFARE enabled) SE chips immediately after the Transaction.

(598) With respect to the second step of the merged entity’s mixed bundling strategy, that is integration of the MIFARE enabled SE within the LTE baseband chipsets and sale of the integrated product in a bundle with the NFC chip, the Commission considers that the merged entity would have the incentive to engage in such conduct.

(599) Respondents to the phase I market investigation stated that the merged entity would have the commercial incentive to pursue a strategy of technical integration in a second step of its mixed bundling strategy. In particular, device OEMs indicated that Qualcomm's business approach so far has been to pursue such integration .

(600) Furthermore, most OEM customers responding to the phase I market investigation declared that they would be interested in purchasing an integrated solution from the merged entity. OEM customers also point out that Qualcomm's integrated solution would make it more difficult for third party NFC/SE providers to compete for customers, as such solution would "lock in" LTE baseband chipset customers . Competitors of baseband chipsets and NFC and SE chips also indicated that device OEMs are likely to be interested in a technically integrated offer .

(601) Furthermore, the Commission notes that the Parties’ internal documents indicate that the Parties have discussed the integration of the SE chip on the Qualcomm baseband chipset, and that each of Qualcomm and NXP expressed a commercial interest to pursue such integration through the Transaction, while maintaining a separate sales channel for the standalone components, consistently with a mixed bundling strategy.

Presentation attached to Email sent by NXP’s […] internally within NXP [DOC ID 1456-28230] [[…] msg].

Qualcomm's internal documents. "[…]", 11 October 2016, [DOC ID 2384-40936 ]

Qualcomm's internal documents, "[…]", 18 March 2017 [DOC ID 2364-52353].

(602) The Commission notes that, in the Article 6(1)(c) decision, based on the responses of the phase I market investigation, it found that the merged entity would likely have the incentive to pursue various types of technical integration, not limited to that of the SE on the Qualcomm baseband chipset, but also that of the NFC chip with the Qualcomm baseband.

(603) However, following the phase II market investigation and on the basis of the review of the Parties’ internal documents discussed in the following recitals, the Commission considers that the merged entity would most likely have the incentive to pursue the specific integration of the SE chip with the Qualcomm baseband chip in the near future, as this is the type of integration that the Parties most extensively discussed in their internal documents.

(604) In the days of the due diligence process leading to the announcement of the Transaction , […], circulated internally within NXP a list of questions received from Qualcomm, concerning security and connectivity (“S&C”). Among the issues raised by Qualcomm appeared questions such as […]. This list of due diligence matters is indicative of the fact that Qualcomm was interested in the potential integration of NFC/SE with its baseband products, and sought to explore this option.

(605) After the announcement of the Transaction, NXP also considered the possibility of technical integration of its products with those of Qualcomm. In an email to […] commented that: […] (emphasis added). The email states that the possibility to integrate the SE within the Qualcomm Snapdragon should be discussed at a later stage down the road within the merged entity. Nevertheless, it is indicative of the fact that NXP internally evaluated this opportunity and had an interest in pursuing product integration.

(606) NXP’s interest in such product integration, in particular of the SE and MIFARE, is reflected in other internal communications. For instance, in an email exchange concerning future discussions to be held with Qualcomm in relation to the security business, […] noted that […] (emphasis added). In reply to this, […] correspondent noted that […] .

(607) The email referred to in recital (606) is also indicative of the fact that not only did the Parties each separately consider and evaluate product integration between the NXP SE and the Qualcomm baseband chipset, but that they also exchanged views on the matter between them.

(608) For instance, in an email to NXP’s […], Qualcomm’s […] commented that […] (emphasis added). In other emails, […] reiterates these views by stating that […] and that […] .

27 October 2016, see https://www.qualcomm.com/news/releases/2016/10/27/qualcomm-acquire-nxp [DOC ID 3346].

NXP internal document, email from […] “[…]”, dated 5 October 2016, [DOC ID 1457-21210] [[…]].

NXP internal document, email from […] “[…]”, dated 5 January 2017, [DOC ID 1456-60344] [[…]].

NXP internal document, email from […] “[…]”, dated 12 January 2017, [DOC ID 1456-12265].

NXP internal document, email from Qualcomm’s […] to NXP’s […], “[…]”, dated 8 February 2017, [DOC ID 1458-35882] [[…]].

NXP internal document, email from Qualcomm’s […] to NXP’s […], forwarding an email by Qualcomm’s […], dated 22 December 2016, [DOC ID 1456-18229].

(609) These emails not only indicate a clear interest on the side of Qualcomm in product integration, but also show that the Parties actually discussed this possibility. Qualcomm in particular viewed the integration of the SE and the addition of MIFARE as a further means to pursue mixed bundling, as the SE would be included […] on the MSM, which would have allowed the sale of a MIFARE-enabled product to a broader set of customers, while maintaining a sales channel for the SE […].

(610) These conversations between the Parties over product integration (and more broadly on the synergies brought by the Transaction) were held, among others, through ad hoc meetings. For instance, on 23 February 2017, NXP’s […] met with Qualcomm’s […]. In preparation of the meeting, […] sent a synergies document entitled […], outlining the main areas of potential synergy between the Parties’ businesses. The document was prepared by NXP’s […], who describes the document as consolidating […] and as explaining […] .

(611) The NXP document mentions, among others, the synergy of […] and […] (emphasis added). Therefore, NXP itself identified integration of the SE as a commercially interesting option, which could be pursued in parallel to the standalone SE, consistently with a mixed bundling approach.

(612) In relation to this synergy, the Commission notes that an earlier draft of the same NXP document had slightly different language, which considered the leveraging of the Parties’ respective market positions more explicitly, and viewed the option of SE integration as more direct […] (emphasis added) . The NXP synergies document’s final summary indicated the potential implication to […]. The aforementioned document suggests that the Parties considered the integration of the SE, which would be sold within the baseband, while a separate sales channel would be maintained.

(613) In commenting internally on the document prepared by […], and […], NXP’s […] commented that the synergies for SMT (“Secure Mobile Transactions”) were […] .

(614) After the meeting of 23 February 2017, Qualcomm’s [...] provided a mark-up to the NXP synergies document, emphasising Qualcomm’s explicit interest in SE integration. Indeed, in the mark-up, […] wrote that […] (emphasis added) .

(615) Qualcomm’s interest in product integration was also manifested in other iterations and meetings held with NXP. In this context, the Commission notes that the Parties’ representatives met at an “integration summit 1”, held in San Diego on 15-16

To that regard, see also the Qualcomm internal discussions held in the email thread entitled “[…]”, 3-5 November 2016, [DOC ID 2364-72772].

NXP internal document, email from […], “[…]” (stating that “[…]”), [DOC ID 1456-54652] [[…]].

NXP internal document, email from […] to Qualcomm’s […], “[…]”, dated 22 February 2017, [DOC ID 1452-11924],

NXP internal document, email from […] to Qualcomm’s […], dated 17 February 2017, [DOC ID 1456-60495].

NXP internal document, “[…]”, pdf document attached to the email from […] to Qualcomm’s […], dated 22 February 2017, [DOC ID 1452-11924].

NXP internal document, email by […], “[…]”, dated 23 February 2017, [DOC ID 1456-18680].

NXP internal document, email from Qualcomm’s […] to NXP’s […], “[…]”, dated 8 February 2017, [DOC ID 1458-35882] [[…]].

NXP internal document, email from Qualcomm’s […] to NXP’s […], forwarding an email by Qualcomm’s […], dated 22 December 2016, [DOC ID 1456-18229].

(619) The first meeting of this security workshop was held on 18-19 April 2017. According to the notes of the meeting, circulated by NXP’s […], […] (emphasis added) .

(620) In a subsequent email, NXP’s […] explained to NXP’s […] that […] (emphasis added) .

(621) The documents discussed in recitals (604) to (620) are indicative of the fact that Qualcomm indeed viewed product integration as the way ahead. Moreover, the Commission notes that NXP also was working on an integration project, Integrated Secure Element Family (“ISEF”), and viewed product integration as the way forward. In this respect, the Commission notes that other NXP internal documents also support such findings.

(622) In relation to Qualcomm’s projects of security integration on the baseband chipset, the NXP notes to a meeting with Qualcomm stated that […]. In commenting on these notes, NXP’s […] noted that […] .

(623) In relation to NXP’s own considerations for SE integration, in an internal presentation entitled […], NXP observed that both Qualcomm and Huawei were developing SEs integrated in the baseband. In an earlier email, in relation to this trend and to Huawei’s SE integration specifically, NXP noted that [… ] .

NXP internal document, email from […] to […], “[…]”, 18 February 2017, [DOC ID 1458-54773] [[…]].

NXP internal document, “[…]”, word document attached to email sent by […] to Qualcomm’s […], dated 15 March 2017, [DOC ID 1458-54816] [[…]].

NXP internal document, email from […] to […], dated 10 March 2017, “[…]”, and attached presentation “[…]”, [DOC ID 1458-55481] [[…]]. It should be noted that, in the email exchange, which included Qualcomm representatives, […] emphasised that the document should be sent back to the business line (“BL)” work stream leads, “[…]”. Therefore, the Parties clearly agreed on the opportunity of integration of the SE into the Qualcomm baseband chipset, and to explore this strategy.

NXP internal document, “[…]”, word document attached to email from […] to […], “[…]”, 3 May 2017, [DOC ID 1456-52124].

NXP internal document, email from […] to […], “[…]”, dated 26 May 2017, [DOC ID 1456-53926].

NXP internal document, email from […] to […], dated 11 May 2017, [DOC ID 1456-51261].

NXP internal document, email from […] to […], dated 3 April 2017, [DOC ID 1458-36014].

NXP internal document, email from […] to […], dated 5 December 2016, [DOC ID 1458-28265] [[…]].

(624) The Commission notes that the Parties actually attributed and calculated synergies related to the product integration. For instance, in a joint presentation dated 27 April 2017, the Parties noted that for Secure Mobile Transactions […] .

(625) Based on the review of the Parties’ internal documents, the Commission therefore notes the following. First, Qualcomm and NXP both had an interest in integrating security features on the Qualcomm LTE baseband chipsets, and had taken steps in that direction. Second, while the integration would require a time frame of one to two years, the Parties considered a transitional short term option, before achieving the integration, of bundled offers. Third, notwithstanding product integration, the Parties’ intention was to maintain a parallel sales channel for standalone products, which is consistent with a mixed bundling strategy. Fourth, the integration and bundling strategy would have included MIFARE, to which the Parties attached value.

(626) This last point is made apparent by the email discussion mentioned in recital (579) above, where Qualcomm executives discussing the synergy importance of MIFARE for mobile transit services commented that […] (emphasis added) .

(627) Finally, the Commission considers that the merged entity would also find it profitable to either raise royalties for MIFARE or, in the extreme, cease licensing this technology to rival NFC and SE chips suppliers altogether, for the following reasons .

(628) First, both competitors and customers of the Parties pointed out that it would be profitable for the merged entity to adopt such a strategy.

(629) Infineon argued that: "There is a very high incentive for QCOM to integrate eSE with MIFARE in basebands/apps processors and worsen the already-prohibitive licensing practices as well as technical maneuvers of NXP" .

(630) […] .

(631) Samsung considers that: "[s]uch a policy would be in line with Qualcomm's current policy against licensing any chipset competitors at all. Given that Qualcomm refuses to license even its FRAND-encumbered SEPs to any chipset supplier, there is little reason to believe that Qualcomm would license MIFARE, which is NXP's proprietary technology even though it has become a de facto standard for at least transit payments".

NXP internal document, “[…]”, slide 2, power point presentation attached to email by […] to NXP’s […], 28 April 2017, [DOC ID 1456-16025] [[…]]. While the document also notes that “[…]”, the Commission notes that this statement acknowledges that the product integration achieved through the Transaction would still produce additional benefits to those pre-existing the Transaction.

Qualcomm internal document, internal email exchange entitled “[…]”, email by […], dated 11 October 2016, [DOC ID 2476-882].

With respect to the possible dissuasive effect of Article 102 TFEU on the merged entity's incentive to engage in an increase in MIFARE royalties or refusal to license MIFARE (See paragraph 46 of the Non-Horizontal Guidelines), the Commission considers that the detection of the conduct at stake under Article 102 TFEU and its legal and economic assessment may be challenging. This is because the assessment of whether a royalty increase or refusal to license MIFARE would be in breach of Article 102 requires a complex analysis on the basis of various ex-post factors. This would make the timely detection of the conduct at stake uncertain. Therefore, it cannot be concluded that the possible applicability of Article 102 TFEU to the conduct at stake would constitute a sufficient deterrent in this case.

See Infineon's reply to Question 3.2. of RFI 29 of 5 July 2017, [DOC ID: 2955].

See response by […] to questions 76 and 77 of Q1 – Questionnaire to device OEMs, [DOC ID: 1009].

See Samsung LSI's response to question 75.1. of Q3- Questionnaire to NFC competitors, [DOC ID: 1095].

(632) Some mobile device OEMs also point out that the merged entity would have the incentive to raise prices for MIFARE-enabled NFC/SE solutions when offered separately from Qualcomm's baseband chipsets. […] explains that: […] .

(633) Qualcomm's internal documents also point to the fact that bundling MIFARE with the SE and LTE baseband chipsets and worsening MIFARE licensing terms of NFC and SE competitors would limit the competitiveness of those providers to Qualcomm's advantage. For instance, in the internal email mentioned in recitals (471) and (579) above, Qualcomm’s […] commented on the implications for Samsung LSI on the lack of a MIFARE licence that […] (emphasis added) .

(634) As regards the Notifying Party's argument that Qualcomm leveraging MIFARE to exclude rival SE chips manufacturers is not merger-specific, as both MIFARE and SE are existing NXP products, and NXP could already engage in a bundling and royalty-increasing conduct today, the Commission considers that such an argument does not hold. First, pre-Transaction NXP does not have the ability to engage in mixed bundling of its MIFARE-enabled SEs (or technically integrate) with baseband chipsets. Second, the incentives of the merged entity to engage in a bundling strategy of these products and degrade the licensing terms for MIFARE are different from those of standalone NXP.

(635) Finally, in the course of its investigation, the Commission received economic submissions from Gemalto and Infineon, which discuss foreclosure concerns associated with bundling by the merged entity of LTE baseband chipsets with NFC/SE chips, in combination with the licensing of MIFARE to rival suppliers of NFC/SEs . Those economic submissions explore three different modelling approaches, referred to as the "Bargaining Model", the "Choi Plus" model, and the

The "bargaining model" studies a setup where an OEM negotiates separately with (i) two competing suppliers of NFC/SE chips (of whom one owns MIFARE, and the other takes a license to this technology), and (ii) with either one or two suppliers of baseband chipsets, depending on the outcome of a stochastic innovation game where either none, one, or both baseband chip suppliers were successful in developing a new generation of baseband chipsets which rendered the old generation obsolete. In this setup, the merged entity would have an incentive to refuse to license MIFARE, as this would give it a competitive advantage in the situation where both baseband suppliers were successful in developing the new generation baseband: In this case, if the OEM had access to an alternative MIFARE licensed NFC/SE supplier (who is not the merged party), the two baseband chipset suppliers would find themselves in neck-and-neck competition for the OEM's custom; but if there is only one available supplier of a MIFARE enabled NFC/SE, namely the merged entity, the latter can also monopolise the baseband chipset market, thus depriving the OEM of the potential benefits of competition in case both baseband suppliers are successful innovators. While this effect would stimulate investments into R&D by the merging party, none of the benefits of these additional investments would be passed on to consumers, who would thus be harmed by such a merger which induces a refusal to license MIFARE. The Commission notes that the predictions of this model appear to be entirely driven by the non-standard assumption that pre-merger the prices for different components are negotiated without taking into account the price of complementary components and that this is changed by the Transaction. The report however does not provide a strong justification or empirical support for why the negotiations on

(636) The "Choi Plus" model submitted by Gemalto/Infineon builds on the general insight that a merger among producers of complementary goods induces the merged entity to lower the price of the bundle composed of its own components, and to raise the price of the stand-alone components. The incentive to offer a discount on the bundle derives from the coordination of pricing decisions of own products post-merger, which tends to lower prices compared to the independent pricing decisions pre-merger (the Cournot effect discussed in recitals (528) to (531) above). At the same time, the merged entity would also have the incentive to increase the price of its standalone components, thus diverting demand away from its rivals' products and towards its own products. The submission received from Gemalto/Infineon assumes in addition that the merged entity would own a must-have IP, namely MIFARE. The rival producers of NFC/SE chips can either license MIFARE from the merged entity before engaging in product market competition with them, or they can develop their own MIFARE compatible technology, which does not practice the merged entity's IP (so that no license to MIFARE is needed) but requires a fixed R&D spend.

(637) According to that model, under certain conditions (in particular if the cost of "inventing around" MIFARE falls within a certain range of values), the merged entity would have an incentive to raise the royalties for MIFARE post-Transaction, which would then translate into an overall increase in prices for all products.

(638) A similar result also occurs when, after the royalty level was determined, but before prices for final goods are chosen, the firms can invest in R&D which improves the quality level of a given component, and hence consumers' willingness-to-pay for any system that contains this particular component. The model shows that the merged entity's incentives to invest would unambiguously increase after the Transaction, while the third parties would reduce their investment, compared to the pre-Transaction scenario, but the net effect on overall investment would be positive. However, even in the presence of such an investment boost, consumers may still be harmed by the impact of the Transaction on royalties and prices, because the merged entity would have less incentive to pass on the benefits of increased innovation to consumers.

one component would pre-merger not be affected by the price for complementary components that are bought by the same buyer and that are needed to produce a given final good.

See also Choi, Jay Pil. "Mergers with bundling in complementary markets." The Journal of Industrial Economics 56.3 (2008): 553-577.

See Non-Horizontal Merger Guidelines, paragraph 117. This pricing externality is thus very similar to the problem of "double marginalisation" in the context of two vertically related firms charging linear prices, and analogously to a conglomerate merger, vertical integration is generally considered as resolving this problem, leading to lower overall margins than those that prevail under vertical separation.

(639) The Notifying Party submitted in response to this model that its results are very sensitive to the underlying parameter values, in particular to the exact R&D cost of "inventing around" MIFARE. Under certain parameter values, the merged entity may either lower (rather than increase) the royalty rate on MIFARE, and under other parameter values, it may even have an incentive to abandon mixed bundling altogether, it will stop offering any discount on its own bundle. This happens when the royalty revenue on MIFARE (which accrue when an NFC/SE rival makes a sale) becomes so important relative to the revenues from sales of its own NFC/SE chips that the merged entity would want to divert demand away from its own NFC/SE chips and towards the rivals' products so as to stimulate royalty revenues.

(640) The Commission considers that the results of the various versions of this model are driven by the assumption that "inventing around" MIFARE is feasible and the cost of inventing around MIFARE being in an intermediate range. The predictions are therefore sensitive to the assumed parameter values. Based on the available information the model cannot be properly calibrated and hence it cannot be reliably predicted whether the merged entity indeed has an incentive to significantly raise the MIFARE royalties. Therefore, the Commission is of the view that this model is inconclusive.

(641) In the light of the above, the Commission considers, notwithstanding the inconclusive economic predictions, based on the overall findings of the phase II market investigation, in particular the Parties’ internal documents, that the merged entity would have the incentive to engage in a strategy of mixed bundling its LTE baseband chipsets with NXP's NFC and MIFARE enabled SE chips (first in a commercial bundle, subsequently in a commercial bundle where the SE has been integrated in the Qualcomm LTE baseband chipset), while at the same time raising the licensing royalties for MIFARE to competing NFC/SE solutions suppliers or, in the extreme refusing to license MIFARE.

3. Likely effects on competition

(642) The Non-Horizontal Merger Guidelines point out that bundling or tying may result in a significant reduction of sales prospects faced by single-component rivals in the market. While the reduction in sales of competitors is not a problem in itself, it may lead to a reduction in the rivals' ability or incentive to compete if the reduction is significant enough, which depends on the characteristics of the industry under consideration. This may allow the merged entity to subsequently acquire or maintain market power. Furthermore, foreclosure practices may also deter entry by possible competitors.

(643) According to the Non-Horizontal Merger Guidelines, in order to be able to foreclose competitors, the merged entity must have a significant degree of market power, which does not necessarily amount to dominance, in one of the markets concerned. The effects of bundling can be expected to be substantial when at least one of the merging parties' products is viewed by many customers as particularly important and there are few relevant alternatives for that product. Further, the Non-Horizontal Merger Guidelines state that for foreclosure to be a potential concern it must be the case that there is a large common pool of customers for the individual products concerned. The more customers tend to buy both products, the more demand for the individual products may be affected through bundling. Such a correspondence in purchasing behaviour is more likely to be significant when the products in question are complementary (emphasis added).

(644) The Non-Horizontal Merger Guidelines further state that the effects on competition need to be assessed in the light of countervailing factors such as the presence of countervailing power or the likelihood of entry which would maintain effective competition.

(645) With regard to the degree of market power held by the merged entity, reference is made to the findings outlined in Sections 7.3.1.2, 7.3.2.2, and 7.3.3.2 for LTE baseband chipsets, NFC and SE chips, and transit service technologies respectively. With regard to the common pool of customers and the incentives for customers to engage in one-stop shopping, reference is made to recitals (544) to (548) above, outlining the existence of a common pool of customers and the advantages for customers to buy a bundled product over a mix-and-match solution.

Effects of mixed bundling

(646) On the basis of its phase II market investigation, the Commission considers that a mixed bundling strategy concerning Qualcomm’s LTE baseband chipsets and NXP’s NFC and SE products (including mixed bundling with the integration of the SE on the baseband chipset), in this case, would appear to be unlikely to lead to foreclosure effects to the requisite standard of the Non-Horizontal Merger Guidelines with regard to providers of baseband chipsets, NFC and SE chips, for the following reasons.

(647) First, the phase II market investigation shows that mobile device OEMs carefully analyse several options on the market before engaging more firmly with one or two suppliers. For instance, […] .

(648) Second, the Commission notes that, notwithstanding Qualcomm’s mixed bundling of chips in past instances (including by way of integration, as referred to in recital (576), device OEMs have not always decided to purchase a bundled product from Qualcomm, rather favouring a standalone product, and have to date obtained from Qualcomm the necessary technical support for mix-and-matching Qualcomm's baseband chipsets with other products, such as the Wi-Fi and Bluetooth chipsets of other suppliers. While those past instances as such do not necessarily indicate that the same would occur in relation to NFC and SE chips, they do suggest that, in this context, a strategy of mixed bundling does not lead to customer device OEMs automatically choosing the mixed bundle, to the detriment of standalone competitors.

(649) For instance, Samsung explained that "In rare cases (Galaxy S8), where Qualcomm’s own offering (say, WiFi) has not met Samsung’s technical requirements, Qualcomm has agreed to provide technical support for a substitution (in this case, with Broadcom’s WiFi chip) and has not “penalized” Samsung by withholding the bundle pricing benefits." Apple also currently mix-and-matches Qualcomm's baseband chipsets with Broadcom's WI-FI and Bluetooth chipsets. […] .

(650) In that respect, the Notifying Party has also submitted data showing that, while its market share and attach rate for WI-FI/Bluetooth chips has increased over time after the acquisition of the WI-FI producer Atheros (see recital (576) above), standalone providers of those chips (notably, Broadcom) are active and have the ability and incentive to compete. According to those data, Qualcomm’s attach rate of WI-FI/Bluetooth chips is low (10%) with respect to high-end OEM customers such as Apple and Samsung. This therefore suggests that OEMs do not necessarily wish to purchase the merged entity’s bundled products and may prefer alternative standalone products.

(651) Third, customer behaviour appears to be inconsistent with foreclosure effects materialising as a result of a mixed bundling strategy of LTE baseband chipsets, NFC, and SE chips.

(652) Apple recently started multisourcing baseband chipsets, as of the iphone7, for which Apple relied on standalone LTE baseband chipsets from Qualcomm and Intel. Apple has confirmed this dual-sourcing approach also for the iphone8, whose LTE baseband chipsets are again supplied by Qualcomm and Intel. With respect to NFC and SE chips, […] .

(653) Apart from Qualcomm, Samsung also can, and does rely on Samsung LSI for the provision of LTE baseband chipsets. In terms of NFC and SE chips, Samsung uses both NXP's and Samsung LSI's products. For the latter case, to deliver a full solution, Samsung LSI produces the NFC chip and partners with Infineon or Gemalto (or both) to deliver the full NFC/SE solution. The Commission also notes that Samsung confirmed that Samsung LSI has also developed a commercially available SE chip (the S3FV9RRX and S3FV9RRP). Therefore, Samsung could rely on its in-house production of LTE baseband chipsets, NFC and SE chips.

(654) Huawei is also self-supplying LTE chipsets. In terms of NFC and SE chips, Huawei sources these from both NXP (NFC and SE chips) and HiSilicon (SE chips).

(655) Therefore, two of the three largest device OEM customers have in-house capabilities for both LTE baseband chips and NFC and SE chips (Samsung with regard to NFC, Huawei with regard to SE). Those options would remain available as possible alternative supply sources to those device OEMs post-Transaction. Apple has at present another supplier of LTE chipset in addition to Qualcomm (Intel).

(656) Fourth, in addition to demand side considerations and customers’ possible counterstrategies, the phase I and phase II market investigation provided strong indications that suppliers of baseband, NFC and SE chips could work together in order to provide similar bundles to that offered by the merged entity (comprising baseband chipset, NFC and SE chips).

(657) Already in the past, where NXP used to be the only provider offering a full NFC/SE solution, other suppliers have worked together to provide a full NFC, SE and SE OS combined solution to mobile device OEMs. For example, in order to provide a full solution to Samsung Mobile, Samsung LSI (baseband and NFC chips supplier) worked with Gemalto (SE OS supplier) and Infineon (SE chips supplier).

(658) These “mix-and-match” options, which device OEMs rely upon today, would remain available after the Transaction. Standalone providers of NFC and SE (such as

See submission of 18 August 2017 by RBB for the Notifying Party, “Assessment of Qualcomm’s selling practices following the acquisition of Atheros”.

See the Notifying Party’s reply to the Commission RFI 51, question 7 [DOC ID 881].

See […] reply to RFI 31, question 2a [DOC ID: 2633], Notifying Party’s reply to RFI 51, questions 9-10, and STMicroelectronics’ reply to RFI 53, questions 1-3, [DOC ID: 3281].

See Samsung’s reply to the Commission’s RFI 28, question 15 c) [DOC ID: 2671].

Gemalto, G&D and Infineon) would maintain the ability and incentive to compete through this mix-and-match approach.

(659) Fifth, in the phase II market investigation the Commission collected information from rival baseband and NFC/SE suppliers regarding the R&D spend, lifetime revenues (including IP revenue, where applicable), lifetime costs (other than R&D) and operating margins for the respondent's five most successful products marketed in the last years. The Commission notes that demand for NFC/SE chips appears to be lumpy, in the sense that if a chip producer wins a socket of a high-volume smartphone, this often entails very large quantities. As regards NFC/SE, based on the feedback to the market investigation, the Commission considers that once a firm is active in the NFC/SE market, even winning one socket of a high-volume smartphone can be sufficient to recoup R&D investments and to achieve minimum viable scale.

(660) Sixth, in relation to the merged entity’s plans to pursue technical integration, the Commission notes that certain market players among chip suppliers (and device OEMs) also have the capability to technically integrate their products, potentially matching the offering of the merged entity.

(661) Recently, MediaTek (which provides baseband chipsets) announced that it was working with STMicroelectronics to integrate its NFC technology on MediaTek's baseband chipset mobile platform. This collaboration, which already won a socket at Nokia, offers a complete mobile payment solution. Additionally, ST Microelectronics also indicated it will start to offer a complete NFC/SE solution.

(662) In addition to the example of MediaTek and STMicroelectronics, the Commission takes note that Samsung also pointed out that, should the market go in the direction of integration, this would not be beyond Samsung LSI and other suppliers' capabilities: "While we cannot say that SLSI (Samsung LSI) or other third parties could not provide an integrated baseband and NFC/SE solution, such a company would have to have the technological base and manufacturing capability to provide such an integrated solution on a cost-competitive basis, and on a reasonable schedule that would permit customers such as Samsung to conduct all necessary testing and certification".

(663) Some of the respondents to the phase II market investigation also pointed out that they would probably work with a baseband chipset supplier to offer a similar bundle or integrated product to that of Qualcomm should competition take place more on bundles or integrated products than on standalone products. For example, Gemalto explained that "If it happens and to position a competitive offer, then we would need to do it by partnering with Qualcomm competition (MediaTek, Hisilicon, Samsung LSI etc) according to whether they can get access to NFC patent pool and MIFARE license to propose such offer". As mentioned above in recital (661), some baseband providers, such as MediaTek, have already entered into such collaboration with NFC/SE chips providers.

See Non-Horizontal Merger Guidelines, paragraph 112.

See http://www.st.com/content/st com/en/about/media-center/press-item.html/t3915 html , [DOC ID 3348].

See minutes of the conference call with St Microelectronics, of 20 June 2017. [DOC ID: 1597].

See Gemalto's response to question 3.3. of RFI 25 of 21 July 2017 [DOC ID: 2152].

(664) The Commission notes that this has been the case also where other chips, such as WI-FI and Bluetooth, were integrated on the Qualcomm baseband chipset, that is, in addition to purchasing standalone solutions, device OEMs could also rely on other suppliers that also developed and offered integrated products. For instance, Samsung explained that it purchases “chipset solutions” from “Qualcomm, S.LSI, Spreadtrum, and Mediatek. Where any part of the chipset solution does not meet the technical requirements of Samsung Mobile, individual chips from Broadcom and/or Qualcomm have been used".

(665) Seventh, the Commission notes that, based on the information provided by the Notifying Party, to the extent that the merged entity would engage in mixed bundling of its baseband chipsets with the NFC and SE, there would still remain a portion of device OEM demand for NFC and SE chips on non-Qualcomm baseband chipsets, (see recital (536) above). This remaining available demand for NFC and SE chips could be addressed by third party providers.

(666) The Commission also considers that technical integration of the SE and baseband chipset in itself would not have significant foreclosure effects, given that in the described scenario of mixed bundling the stand-alone components would be still available. To the extent customers decide to buy the integrated product despite the option to mix-and-match with components of the merged entity or of third party providers, this could be due to advantages implied by technical integration.

(667) Therefore, for the reasons outlined above, the Commission takes the view that baseband, NFC and SE competitors of the merged entity would likely not be foreclosed, even if the merged entity were to offer a bundled (and subsequently integrated) offer of LTE baseband chipsets with NFC and SE chips at a lower price than the standalone components, and would maintain the ability and incentive to compete. This finding would be independent of whether the SE chip in the bundle (initially in a commercial bundle, subsequently integrated with the baseband chipset) would be MIFARE-enabled or not.

Effects of increased royalties for MIFARE in addition to mixed bundling

(668) Notwithstanding the assessment in recitals (646) to (667) above, the Commission also considers that, as shown in Sections 1 and 2 above, the merged entity has the ability and incentive to not only equip its own products with MIFARE, but also to raise royalties for MIFARE to competing NFC and SE suppliers (or cease licensing of MIFARE altogether).

(669) As explained in Section 7.3.3.2 above, with MIFARE the merged entity would have a dominant position in the market for transit service technologies, as MIFARE is the leading transit service technology and is a particularly important and required technology by device OEMs.

(670) In this context, respondents to the phase II market investigation emphasised the importance of MIFARE, and argued that an increase in royalty fees for MIFARE or a refusal to license MIFARE to standalone NFC and SE producers would seriously impact the competitiveness of their products in the near future, when device OEMs are expected to require all SEs to be MIFARE certified (this has been mandatory only for high-end phones until recently).

See Samsung’s reply to question 19g of Commission RFI 28 to Samsung, [DOC ID: 2671].

(671) Almost all NFC and SE chipset competitors considered that, if they were unable to offer MIFARE-enabled products because of increased royalties or a refusal to license MIFARE, it is unlikely that OEM customers would consider them as an alternative to the merged entity’s products.

(672) Gemalto argued that should Qualcomm bundle or tie MIFARE to its LTE baseband chipsets, and cease the licensing of MIFARE to competitors, then "in both cases the result would be a foreclosure for both NFC/SE solutions and baseband/AP chipsets, MIFARE being a mandatory features in more and more smartphones".

(673) Infineon argued that: "The potential impact of anti-competitive behaviours have a high level with MIFARE. QCOM will be uniquely positioned to benefit from this widely-deployed but proprietary technology – leaving out other industry players who do not have a MIFARE license. The anti-competitive effects arise from the fact that smartphone OEMs will have a very limited set of suppliers to choose from while implementing mobile ticketing in their smartphone models."

See for instance response by Gemalto to question 76 of Q3 – Questionnaire to NFC competitors, [DOC ID: 678].

(674) Device OEMs confirmed this view. […]. Samsung also pointed out that MIFARE is a mandatory requirement for its smartphones .

(675) Furthermore, it appears that the MIFARE requirement may not be circumvented by offering "MIFARE compatible" SEs. When asked whether they could develop and offer "MIFARE-compatible" products without a MIFARE license from the merged entity, suppliers of NFC and SE chips pointed out that they would encounter obstacles that they consider almost impossible to overcome.

(676) Infineon explained that one of those obstacles are the MIFARE trademarks, and that inventing around the MIFARE trademarks is impossible; another major obstacle would be certification: even in the hypothetical event of successful development of a MIFARE compliant eSE without MIFARE license, competitors believe that it would be impossible to get a MIFARE certification as the certification bodies like UL or Arsenal would not be willing to certify products without NXP license. Infineon pointed out that in any case any "MIFARE compatible" product would have a lower value for the customers than a licensed product, even if it could be called "MIFARE compatible". Customers insist on certifications, while such unlicensed solutions are unlikely to be certified by the certification partners of NXP .

(677) Gemalto also pointed out another major obstacle in case access to MIFARE cannot be obtained, which are the expected costs to develop MIFARE compliant products which do not infringe NXP's MIFARE IP .

(678) Therefore, should the merged entity, in addition to bundling baseband, NFC and (MIFARE-enabled) SE, raise MIFARE royalties to competitors or cease the licensing of MIFARE altogether, this would change the competitive conditions in the markets for baseband chipsets and NFC and SE chips.

See for instance response by Gemalto to question 76 of Q3 – Questionnaire to NFC competitors, [DOC ID: 678].

(679) In particular, the Commission considers that, as a consequence of this conduct, competitors of the merged entity would not be able to react to the merged entity by offering a bundle comprising MIFARE-enabled SE or would only be able to offer it at unattractive prices compared to that of the merged entity.

(680) An increase in MIFARE royalties would be likely to (i) directly raise rivals' costs in the NFC/SE segment (Gemalto, Infineon) because a crucial input for these rivals, namely the MIFARE license, would become more expensive, and (ii) divert demand away from rival baseband chipset (MediaTek, Intel) suppliers, because the complementary components to these basebands, the MIFARE enabled NFC/SE chips (both the merged entity's and the rival NFC/SE suppliers' products), would become more expensive.

(681) In the extreme, should Qualcomm enable MIFARE on its own SEs and deny it (or license it under terms which make it unprofitable) to other competitors, competing NFC/SE suppliers could not offer MIFARE enabled NFC/SE chips any more. In that case, mobile device OEMs would prefer the Qualcomm offering, given the importance of the MIFARE requirement for the mobile device OEMs, which would make the Qualcomm bundle more attractive. In the absence of alternative MIFARE-licensed products, supplier options for device OEM customers, and their ability to mix-and-match the products of the Notifying Party with those of other suppliers, would be more limited than in a scenario where MIFARE remains accessible at the current terms to competing NFC/SE producers.

(682) The expected anti-competitive effects of deteriorating the licensing terms of MIFARE (or withholding it altogether) appear to be particularly pronounced because alternative transit service technologies appear to be poor substitutes from the point of view of mobile OEMs (see section 7.3.3.2) as is evident from the requirement often imposed by mobile OEMs for NFC/SE products to be MIFARE-enabled. Therefore, if the access terms for MIFARE of competing NFC/SE chip producers deteriorate, this has a direct negative effect on their ability to compete with the merged entity on the relevant NFC/SE market.

(683) The Commission notes in this context that the economic submission referred to as the "Choi Plus" Model also concludes that, to the extent that the Transaction leads to (i) mixed bundling by the merged entity, and (ii) an increase in the royalty rates for MIFARE, rival producers of both NFC/SE chips and of LTE baseband chips are unambiguously weakened .

(684) Even if the royalty increase on MIFARE were not prohibitive, so that rival NFC/SE producers would still be viable even under the new level of royalties, such an increase would still have a potentially negative impact, because the MIFARE royalties represent a variable cost to third party producers, which will therefore be at least partially passed on to mobile OEMs through higher product prices. As a result, the competitive constraint exerted by third parties on the merged entity would diminish.

(685) Further, the Commission considers that, as a result of this conduct of increased MIFARE royalties (or refusal to license) and mixed bundling, the profitability of the merged entity’s competitors is likely to decrease. Those competitors may thus find it more difficult to invest in the further development of their products. In this respect,

Report submitted by CRA titled "Economic Modelling Confirms Risks to Competitors and Consumers" submitted on 14 August 2017. [DOC ID 2936]

the phase II market investigation provided strong indications that the markets for LTE baseband chipsets, NFC and SE chips and MIFARE are R&D intensive and upfront substantial investments in the development of the products are essential to remain competitive on these markets. Eventually, lower incentives to invest in R&D may therefore weaken the competitive constraint imposed by the Notifying Party's rivals, in the foreseeable future .

(686) Gemalto explained that the market for NFC, SE and SE OS are R&D expenditure driven: "NFC front end can now be considered as a stable technology as most of the functions are standardized (RF, NCI interface to AP etc) and follow a type approval process. But it still requires some level of R&D investments as with any silicon component requiring to be redesigned in more and more advanced nodes. The need for continuing R&D is even more the case with the eSE IC (HW silicon) as it also follows the regular silicon life cycle towards more advanced nodes (a new node is introduced about every two years) but also requires regular security improvements. For the eSE OS, the objective is to contribute to the OEM competitiveness, differentiating on its secure services strategy thanks to the eSE OS. As such, first investment is to integrate with the selected NFC controller front end, then raise the OS to support existing secure service, integrate new ones according to an agreed upon roadmap and do porting on new eSE IC technology to remain competitive. This is for the sole part of the eSE product as in addition, it must be accompanied by support for integration on the different phone models and service set as well as continuous quality monitoring and defect analysis. This is to be multiplied by the number of phone models and is cumulative over the years. This is an R&D expenditure driven market". Infineon also argued that: "The majority (>90%) of the R&D expenses are typically invested before any device is selected by an OEM, which includes agreeing on the unit price" .

(687) G&D also stated that: "In G+D’s view the products continuously evolve to provide better user experience in regards of performance, security or supporting new use cases as well as cost and space optimizations for OEMs or to cope new requirements from standardisation bodies. G+D considers these investments as significant since it comprises e.g. development of new technologies for the NFC controller (e.g. booster for the RF), functional and technological enhancements of the eSE. The SE OS needs to reflect new HW, security and use case requirements which results in high efforts to further optimize performance and implement new features. For the latter two expenditures for the certification of the respective products is due".

(688) Therefore, before committing to significant R&D expenditures, suppliers would need to ensure that their product or bundle responds to at least the mandatory technical requirements requested by mobile device OEMs. One of these requirements is MIFARE. As a consequence of the Transaction, NFC/SE suppliers would either have to face increased costs (in case of an increase in MIFARE royalties) or would not be able any more to comply any more with this requirement (in case the merged entity stops licensing MIFARE).

See question 31 of Questionnaire 11 to Competitors of NFC and question 1.14 of RFI 25 to Gemalto of 21 July 2017[Doc ID: 2152], question 1.13.3 and 1.16 of RFI 29 to Infineon of 5 July 2017 [DOC ID: 2955] and question 1.14 of RFI 26 to G&D of 4 July 2017 [DOC ID: 1932].

(689) Based on the results of the market investigation, the Commission therefore takes the view that increasing royalties for MIFARE or ceasing the licensing of MIFARE altogether to competitors against the backdrop of the merged entity's strategy of mixed bundling of LTE baseband chipsets, NFC SE chips and MIFARE would have the effect of foreclosing competitors of baseband chipsets and NFC and SE chips who would not be able to engage in timely counterstrategies and overcome obstacles related to the more restrictive conditions regarding the licensing of MIFARE.

7.4.3. Pure bundling and tying

7.4.3.1. Background

(690) In this section, the Commission will assess whether the merged entity would engage in a strategy to foreclose competitors through pure bundling or (commercial/technical) tying of its LTE baseband chipsets with NXP’s NFC and SE products.

(691) At the outset, the Commission recalls that, based on the results of the phase II market investigation, illustrated in Sections 7.3.1.2, 7.3.2.2 and 7.3.3.2 above, the Commission considers that the merged entity would hold a dominant position in the worldwide market for LTE baseband chipsets and in the market for transit service technologies, as well as a certain degree of market power in the markets for NFC and SE chips. Furthermore, these products are particularly important for device OEM customers and they are complementary products, purchased by a common pool of customers.

(692) With regard to pure bundling, the merged entity’s products would be only sold jointly in fixed proportions, and not available separately. For the assessment of the present case, in a pure bundling scenario the merged entity would no longer offer standalone LTE baseband chipsets and NFC, SE chips (enabled with MIFARE) standalone, but only together in a bundle comprising those products.

(693) With regards to tying, as explained in Section 7.1 and 7.4 above, according to the Non-horizontal Merger Guidelines, tying refers to a situation where customers that purchase one good (the tying good) are also required to purchase another good (the tied good). Tying can take place on a technical or commercial (contractual) basis.

(694) For instance, technical tying occurs when the tying product is designed in such a way that it only works with the tied product (and not with alternatives offered by competitors). Contractual tying entails that the customer when purchasing the tying good undertakes only to purchase the tied product (and not the alternatives offered by competitors) .

(695) In the phase I market investigation, the Commission investigated whether the Notifying Party would have the ability and incentive to engage in a foreclosure strategy by either: (i) technically integrating NXP's NFC and/or MIFARE-enabled SE chips to Qualcomm's LTE baseband chips and no longer providing these products standalone; (ii) commercially tying NXP's NFC and/or MIFARE-enabled SE chips to Qualcomm's LTE baseband chips and no longer making available to customers LTE baseband chipsets standalone and (iii) commercially tying Qualcomm's LTE baseband chipsets to NXP's NFC and/or MIFARE-enabled SE chips and no longer making available those NXP chips standalone.

Non-Horizontal Merger Guidelines, paragraph 97.

(696) The Commission will analyse the potential conducts of pure bundling and tying together referred to in the following Section, given the similarities between the two possible behaviours, which both presuppose that the merged entity would commercially or technically bind its products, foregoing the sale of the relevant standalone components.

(697) With regard to a possible tying or pure bundling conduct that would be inclusive of MIFARE (that is to say, MIFARE would only be available in a pure bundle or with a tying product), the analysis remains the same irrespective of whether the SE chip would be MIFARE-enabled. The Commission notes that, should the merged entity also withhold MIFARE entirely by refusing to license it, such conduct has already been analysed under Section 7.4.2.2 above. Moreover, the commitments submitted by the Notifying Party to address the competition concerns identified by the Commission in relation to mixed bundling together with increased royalties of MIFARE, also remove any possible competition concerns arising from a conduct of tying or pure bundling that would include an increase of royalties or refusal to license MIFARE.

7.4.3.2. Analysis of pure bundling and tying

(698) Following the phase II market investigation, the Commission considers that the merged entity would lack the incentive to engage in the conducts of pure bundling and tying of LTE baseband chipsets and NFC/SE chips.

(699) Moreover, should the merged entity have the incentive to engage in a conduct of pure bundling or tying, the Commission notes that in any event such conducts, if implemented, would be unlikely to lead to significant foreclosure effects.

A) The Notifying Party's view

1. Ability

(700) With regard to pure bundling, the Notifying Party reiterates the arguments made with respect to mixed bundling illustrated in Section 7.4.2.2 A) above, that is to say, the Parties do not have market power with respect to LTE baseband chipsets or NFC and SE chips, their products do not have a “must have” status, and device OEMs exert countervailing buyer power and can rely on alternative product offerings from competing suppliers of standalone LTE baseband chipsets and NFC/SE chips.

(701) With respect to the ability to coerce customers into accepting a pure bundle of LTE baseband chipset with an NFC and SE solution (and MIFARE), the Notifying Party submits that it is not plausible to suppose that a customer mobile device OEM would base its choice of baseband chipsets on the specific NFC or SE chip that may be offered with that baseband chipset. NFC and SE chips support peripheral systems that are "good to have" rather than "must have" inputs like baseband chipsets. Given the importance of baseband chipsets, a customer would make its choice on the basis of the baseband chipset, regardless of the inclusion of NFC and SE in a pure bundle. The Notifying Party also claims that a pure commercial bundle would be the equivalent of a price increase of all units, affecting a substantial part of the demand of those device OEMs that do not wish to incorporate NFC functionality in their mobile devices. Such price increase would cause the diversion of substantial sales to rival chip suppliers selling standalone baseband chipset.

(702) Moreover, the Notifying Party submits that there are also practical impediments to practices such as pure bundling or tying: baseband chipsets and NFC and/or SE are purchased through different procurement teams, and at different times. At a minimum, the fact that customers do not currently structure their tendering processes

134

in a manner compatible with BC and NFC/SE bundling poses an obstacle to the use of a bundling strategy by the merged entity and serves to make such a strategy less viable .

(703) With regard to Apple specifically, the Notifying Party considers that it would not have the ability to coerce Apple to accept a bundled offer for the following reasons: (i) Apple intends to continue using an NFC/SE combined solution chip in its future mobile devices, and more specifically one that resides on a single chip (according to NXP's roadmap, NXP plans to develop such a combined chip for Apple); and (ii) Apple alone accounted for 67% of the revenues NXP derived from its mobile transactions product line in 2016 (which includes NFC/SE sales), therefore a decision on Apple's part to switch away from NXP would mean that its share of supply would decrease substantially.

(704) With regard to Samsung, the Notifying Party submits that it would not have the ability to coerce Samsung to accept a bundled offer either for the following reasons: (i) Samsung currently dual-sources integrated baseband chipsets from the Notifying Party and from its own subsidiary, Samsung LSI, and in the past has chosen to use both non-integrated and integrated BC solutions; and (ii) Samsung has several alternatives to NXP's secure mobile payment solutions at its disposal: it has developed Samsung Pay which relies on NFC, HCE and Samsung's own MST technology. Samsung LSI has also developed its own NFC chip which is now used in Samsung smartphones.

(705) With respect to tying, the Notifying Party submits that it does not have the ability to engage into a commercial or technical tying (that is to say integration of NXP's NFC and SE chips into the Notifying Party's baseband chipsets) strategy with respect to Qualcomm's LTE baseband chipsets and NXP's NFC/SE chips with the purpose of foreclosing its competitors.

(706) In relation to the possible integration of the NFC controller into the MSM and/ or MDM, the Notifying Party claims that it does not have extant such plans because: (i) the current and expected adoption rate of the NFC functionality is insufficient to justify the investment in such integration work; (ii) the average price of NXP's NFC controllers is merely between […] to […] and (iii) the integration of the NFC functionality would carry certain integration risks, namely the complexity of the certification process relative to other technologies (Wi-Fi, Bluetooth, FM) that could negatively impact the time to market of an integrated solution.

(707) Further, as regards the possible integration of the SE into the MSM and/or the MDM, the Notifying Party argues that it has already partially integrated SE features in one of its MSMs and it is indeed considering incorporating a fully-functioning SE chip into the application processor part of the MSM. The full integration would most likely occur in the near future, […], because it would require developing the requisite software to run on Qualcomm's hardware, and obtaining carrier aggregation.

(708) The Notifying Party also submits that it would integrate the SE in its MDMs provided that there is a business case for this. As Apple is the main user of the Notifying Party's MDMs and by far the main purchaser of NXP's SE and Apple only sources standalone BCs, there does not appear to be such business case.

(709) As regards the possible integration of the NFC/SE combined solution chip into MSM and/or MDM, the Notifying Party considers that such integration would be

620 RBB economics, Response to theories of harm relating to product bundling and tying. [DOC ID: 1745].

135

challenging and submits that currently it has no current plans to do so. Even if the Notifying Party were able to technically integrate NXP's secure mobile payment products into the baseband chipset at some point, integration would only concern NXP's SE chips, and commercial penetration of such a solution would be very limited. Therefore it would be in the Notifying Party's financial interest to continue offering those products on a standalone basis, and in the foreseeable future, sales by the merged entity of standalone NFC and SE chips would remain far larger than sales

621 of integrated products .

(710) Even if the merged entity were to proceed with integration of the SE into the MSM and/or MDM, the Notifying Party submits that such technical integration would occur only in respect of newly developed and released baseband chipsets and could therefore only meet the requirements of a certain portion of demand. The Notifying Party estimates that any hypothetical technical integration would be incorporated in […] of those Qualcomm baseband chipsets sold in 2019-2020 .

2. Incentive

(711) As regards pure bundling, in terms of incentive, according to the Notifying Party, bundling would make no commercial sense as NFC/SE chips are not considered "must have". Contrary to baseband chipsets that drive the core functionality of the smartphone, NFC and SE chips are not essential to the functionality of mobile devices. They only have a high penetration as regards high-end phones and the NFC/SE functionality can be achieved through the use of a variety of rival technologies. Qualcomm would thus be unable to leverage NXP's position in NFC and SE chips to coerce OEMs into buying baseband chipsets. This would greatly limit the merged entity’s incentive to offer those products in a pure bundle.

(712) Furthermore, the Notifying Party submits that the price differences of baseband chipsets and NFC and SE chips render pure bundling (or commercial tying) unlikely. The Notifying Party explains that the profitability of a pure bundling strategy depends on two key factors: the profit margins earned on each product in the bundle and the reactions of affected customers. More formally, the relevant profit margins will imply a “critical switching level”, the percentage of Qualcomm baseband chipsets sales that would need to be lost as a result of the tying strategy to render that strategy unprofitable. Evidence on how affected consumers would react then allows us to determine whether “actual switching levels” can be expected to exceed this critical level, and as such whether this strategy is indeed likely to be unprofitable.

(713) According to the Notifying Party, NXP earns an average margin of […] per unit on standalone NFC sales and […] per unit on sales of NFC/SE combinations. Further, NXP estimates that […] of rival NFC sales are of standalone NFC units, while […] are sales of NFC/SE combinations. On that basis, and assuming any customers who switch to NXP would utilise standalone NFC and NFC/SE combinations in similar proportions, the merged entity might be expected to earn an average margin of around […] per unit after the Transaction on any additional NFC sales captured as a result of the tie. In comparison, Qualcomm currently earns an average BC margin of […] on supply to its eight largest OEM customers.

621 Form CO, paragraphs 891 and following [DOC ID 326].

622 See response by the Notifying Party of 28 June 2017 to the Article 6(1)(c) decision, paragraph 84, [DOC ID 1331].

(714) Taking the overall weighted average, the critical switching level is […], equal to […]. Thus, if more than […] of relevant purchases of Qualcomm BC that would be combined with non-NXP NFC absent the tie would switch to other sources of BC in response to the tie, then the tying strategy would be unprofitable .

(715) Therefore, even if a small percentage of customers switched away from the merged entity's baseband chipsets, that would be enough to render the merged entity pure bundling strategy unprofitable.

(716) Then, a foreclosure strategy based on bundling or tying of Qualcomm’s BCs with NXP’s NFC and/or SE solutions (or vice versa) would not be profitable for the merged entity as certain key customers who would otherwise have chosen to purchase one of the products from the merged entity on a standalone basis may switch to the merged entity’s rivals in response to such a hypothetical tying/bundling foreclosure strategy. Given that the two main purchasers of NXP’s NFC chips are Apple and Samsung, with the former exclusively using its own AP, and the latter having in-house baseband chipset, application processor and NFC capabilities, a hypothetical tying strategy by the merged entity aimed at foreclosing rivals from supplying these extremely powerful customers would most certainly fail, and would instead lead them to shift their purchases of NFC products away from the merged entity, in part or in full. In such a scenario, the merged entity would forego significant profits.

(717) In particular in the case of a hypothetical foreclosure strategy whereby NXP’s NFC/SE is the tied product, it would be in the interest of the merged entity to encourage the widest possible adoption of NFC/SE technology, given its use in mobile devices is still at an incipient stage and that it faces competition from a range of alternative technologies that could displace it at any moment. Therefore, the merged entity will have an incentive to continue to market NFC chips on a standalone basis, as tying them to BCs is likely to restrict adoption of NFC technology.

3. Effects

(718) The Notifying Party submits that a pure bundling conduct would not lead to any foreclosure effects, as device OEMs are sophisticated buyers, which can and are willing to source from multiple available options. In that regard, the Notifying Party refers to the “mix-and-match” solutions that are available from standalone suppliers of NFC and SE chips. The Notifying Party further notes that there will be an increasing amount of non-Qualcomm baseband chipsets in the coming years, that would be available to third party NFC and SE suppliers, which thus would not be foreclosed by a pure bundling conduct (see recital (536) above). The Notifying Party claims that it has not previously impaired competition in other markets adjacent to baseband chipsets supply, such as Wi-Fi and Bluetooth chips after the acquisition of Atheros in 2011 .

(719) As regards tying, the Notifying Party submits that even assuming that the merged entity would have the ability and incentive to engage in a foreclosure strategy through the tying of Qualcomm’s BCs with NXP’s NFC and MIFARE-enabled SE chips (or vice versa), such practices would have no anticompetitive effects on either of those product areas.

(720) A hypothetical bundling or tying strategy by the merged entity would not significantly reduce the incentives and ability to compete of single-component rivals. In the Notifying Party's view, large processor manufacturers such as MediaTek will continue to sell best-of-breed baseband chipsets on a standalone basis, in direct competition with the merged entity. Similarly, suppliers such as STMicroelectronics will continue to sell NFC and SE products on a standalone basis due to the purchasing strategies of large customers such as Samsung and Apple. In addition, large chipset suppliers such as Samsung LSI will continue innovating and offering NFC and SE chips, and baseband chipsets, both on a standalone and integrated basis.

(721) Furthermore, device OEM customers have sufficient countervailing buyer power to prevent the combined company from attempting any foreclosure strategy. Customers value cross-technology interoperability and the ability to mix-and-match solutions from different suppliers so as to better meet their technical requirements and differentiate their products. Customers will therefore prevent suppliers from attempting to put in place bundling strategies that do not enable them to realise their commercial strategy.

(722) For those reasons, in the Notifying Party's view the Transaction would not result in anticompetitive effects from a possible combination of Qualcomm’s BCs with NXP’s NFC and/or SE.

B) Commission's assessment

1. Ability to engage in pure bundling or tying

(723) With respect to the merged entity’s ability to engage in pure bundling or tying of LTE baseband chipsets, NFC and SE chips, and MIFARE, the Commission notes the following.

(724) First, the Commission recalls that baseband chipsets represent an indispensable component for any mobile device and NFC/SE technology is expected to increase in importance and attach rate in the next two to three years (see recital (872) below).

(725) Second, as mentioned in recitals (546) to (548) to above, LTE baseband chipsets, NFC and SE chips, and MIFARE are complementary products and are purchased by the same customers, namely device OEMs.

(726) Third, with regard to the Notifying Party’s argument that a pure bundling or tying conduct could not be implemented, given that device OEMs have different procurement teams operating at different times, the Commission recalls that, as explained in recitals above (554) to (555) and (567) to (569) above, the merged entity would be able to sell the products jointly to device OEMs, and indeed Qualcomm’s plan was to combine the Parties’ sales teams to offer products to device OEMs under the same sales account. Therefore, device OEMs’ procurement processes would not be an impediment to a pure bundling conduct.

(727) Fourth, when asked whether Qualcomm would have the ability to engage in pure bundling, namely to offer its baseband chipsets exclusively together with NXP's NFC/SE and no longer making those products available separately to customers,

138

device OEM customers, and baseband chipset and NFC competitors responding to the phase I market investigation indicate that Qualcomm has that ability .

626

(728) For instance, […] .

(729) In that regard, the Commission notes that there would be no technical impediments for the merged entity to sell LTE baseband chipsets and NFC and SE chips bundled together to device OEMs. This approach of selling products as a “platform” is also consistent with Qualcomm's past behaviour. For instance, Samsung explained that “Qualcomm offers “chipset solutions” or “mobile platforms” that consist of a MSM (which is usually a combination of the AP and modem) and other chips that comprise RFIC (Radio Frequency IC), PMIC (Power Management IC), WiFi, BT and GPS, among others. Sometimes (but not always) BT and WiFi are combined into a single chip, as are the RFIC and GPS […] For technical reasons, bundling permits faster deployment (as all of the chips in this solution are already integrated) and arguably has pricing advantages (as less time needs to be spent with integration of different solutions). The technical reasons for bundling also conveniently permit the chipset solution provider to price its chipset as a bundle, whether for better or for worse” . Samsung further explained that “the components within the Qualcomm integrated solution are optimized to work together and are priced as a platform” .

(730) Fifth, despite the Notifying Party’s assurances that the merged entity would not proceed in technical integration, the Commission recalls that, as illustrated in recitals (557) to (579), based on the results of the phase I and phase II market investigations and the analysis of the internal documents of the Parties, the merged entity would have the ability to technically integrate NXP's MIFARE-enabled SE chip on Qualcomm's LTE baseband chipset. The vast majority of the respondents to the phase I and phase II market investigations consider that Qualcomm has the technical know-how and resources to implement such a technical integration of NFC and SE chips on the baseband chipset.

(731) Furthermore, some of the respondents pointed out that Qualcomm has a history of integrating various components onto its Snapdragon SoC and provided such examples. As shown in recitals (601) to (626), the Parties' internal documents provide proof that Qualcomm has already extant plans with regard to integrating NXP's NFC and MIFARE – enabled SE chips on Qualcomm's baseband chipsets.

(732) The Commission further notes that, as mentioned in recital (576) above, in previous instances, Qualcomm has proceeded to integrate other components, such as Wi-Fi and Bluetooth, on its baseband chipset, and to sell those products together. However, in those past instances, Qualcomm has also made available standalone solutions for customers .

(733) Sixth, as regards the merged entity's ability to engage into a strategy of commercial tying, the Commission takes the view that the merged entity would have the ability to

625 See responses to question 70 of Q1 – Questionnaire to device OEMs; responses to question 62 of Q2 – Questionnaire to baseband competitors; responses to question 68 of Q3 – Questionnaire to NFC competitors.

626 See response by […] to question of 57.1. of Q1 – Questionnaire to device OEMs, [DOC ID: 1009]

627 See response by Samsung to question 19 of RFI 28 [DOC ID: 2671].

628 See response by Samsung to question 7 of the Commission’s supplemental questions to RFI 28 [DOC ID: 2672].

629 RBB Economics paper of 18 August 2017, "Assessment of Qualcomm's selling practices following the acquisition of Atheros", pp 9-10.

139

engage in such a strategy. In fact, in implementing such commercial tying strategy immediately after the Transaction, the merged entity would not encounter any technological hurdles, and the products involved are purchased by common pool of devices OEM customers, which facilitates the implementation of a commercial tying strategy. Furthermore, respondents to the phase I market investigation consider that the merged entity would have the ability to engage in such strategy .

(734) Overall, based on the above the Commission takes the view that the merged entity would have the ability to engage in possible conducts of pure bundling or tying of LTE baseband chipsets, NFC and SE chips.

2. Incentive to engage in pure bundling or tying

(735) As the Non-Horizontal Merger Guidelines point out, the merged entity's incentive to engage in tying or bundling depends on the degree to which such strategy is profitable for the merged entity .

(736) When asked whether Qualcomm would have the commercial incentive to engage in pure bundling, namely by offering its baseband chipsets exclusively together with NXP's NFC/SE and no longer making those products available separately to customers, most respondents to the phase I investigation indicated that Qualcomm has that incentive . […] .

(737) During the phase I market investigation, respondents also indicated that Qualcomm would have the incentive to engage into a strategy of technical or commercial tying of Qualcomm's LTE baseband chipsets with NXP's NFC and MIFARE-enabled SEs with the purpose of foreclosing its competitors in both markets .

(738) One mobile OEM argued that the merged entity would find it profitable to engage in commercial tying whereby it would tie Qualcomm's LTE baseband chipset (tied product) to NXP's NFC and MIFARE-enabled SE chips to its LTE baseband chipset (tying products). The respondent argues that taking into account the respective prices of baseband chipsets and NFC and SE chips, even if a small proportion of NXP's existing customers who do not purchase Qualcomm baseband chipsets today were to switch to Qualcomm as a result of the commercial tying strategy, such strategy would be profitable.

(739) The Commission considers that (especially compared to engaging in mixed bundling) pure bundling or tying often entails to sacrifice some profitable sales of standalone components the merged entity would otherwise have made. On the other hand, especially when competitors are foreclosed, the merged entity may be able to raise price in the concerned market(s) due to increased market power. The merged entity thus faces a trade-off between the possible costs associated with bundling or tying its products and the possible gains from foreclosing competitors in the concerned market(s). The incentives to engage in bundling/tying therefore depend on the prospect of being able to foreclose competitors.

(740) For the reasons set out in recitals (761) to (766), the Commission however considers that even in the event that the merged entity pursued a bundling/tying strategy, foreclosure effects would be unlikely. Therefore, the possible gains for the merged entity from implementing a bundling/tying strategy with a view to foreclosing competitors are limited.

(741) The Commission further notes that, while NFC and SE chips are important for smartphones and are likely to become even more important in the future, the choice of the baseband chipset vendor is a much more significant decision for a device OEM customer than the choice of its NFC and SE provider. Therefore the Commission takes the view that device OEM customers are unlikely to make their choice of baseband chipset provider dependent on the choice of the NFC and/or SE chipset provider.

(742) As regards the Notifying Party's argument that even if a small percentage of customers switched away from the merged entity's baseband chipset, that would be enough to render the merged entity’s tying or pure bundling strategy unprofitable (see recitals (712) to (715) above), the Commission notes that this argument is flawed for at least two reasons.

(743) First, the Notifying Party's calculations are based on the assumption that post-Transaction, the margin on the bundled product will remain unchanged relative to the pre-merger level of the joint margins of the separately sold products, which is unlikely in particular if the pure bundling strategy were indeed effective at foreclosing rival suppliers. In this case, the merged entity could profitably increase the price for its bundled product, thus leading to higher margins than prevailed pre-merger.

(744) Second, the Notifying Party limits itself to explaining the costs of pure bundling in terms of business lost on the baseband chipset side; but if there is indeed such a strong difference between LTE baseband chipsets and NFC/SE chips in terms of price and importance for the customer as claimed by the Notifying Party, then it is likely that only very few, if any, customers will switch away from the merged entity's LTE baseband chipsets just because they are obliged to buy the merged entity's NFC/SE chip as well.

(745) Those considerations suggest that the incentives of the merged entity to engage in bundling/tying are rather limited, as this would entail the prospect of losing profits from not serving customers that want to mix-and-match anymore, whereas

636 Ceasing to supply stand-alone products may also be profitable for the merged entity (even when ignoring potential gains from foreclosing competitors) if maintaining the stand-alone products would entail large additional fixed costs, for example for R&D or marketing.

637 Compared to mixed bundling, depending on the circumstances, pure bundling/tying may allow the merged entity to divert more demand away from competitors compared to mixed bundling, so that potentially there may be a more pronounced foreclosure effect.

638 See Non-Horizontal Merger Guidelines, paragraph 105.

141

foreclosure would be difficult to achieve. In the course of its investigation, one of the economic submissions received by Gemalto/Infineon, namely the "Choi and Stefanadis Plus" model, outlines why merged entity could have an incentive to engage in pure bundling. The model relies on the idea that rival suppliers of either LTE baseband chipsets or NFC/SEs in light of the weaker market position compared to Qualcomm/NXP have to invest into product development first, before being able to compete (with possibly better products) against Qualcomm/NXP. After the Transaction, the merged entity may have an incentive to technically integrate its products so that the potential rivals can recover their investment costs only in the (less likely) case that both of them innovate successfully, but not when only one of them is successful (in which case the innovator can only make money if mix-and-match with the complementary product of the merged entity was possible, but this is exactly what is prevented by technical integration); this makes the rivals' investment riskier and less profitable, thus discouraging rival innovations and reducing competitive pressure on prices.

(746) According to the model, whenever this bundling strategy is profitable, it is also harmful to consumers. The model also shows that the Transaction may be harmful to consumers even absent bundling; this follows from the fact that the Transaction allows the merged entity to coordinate the investment decisions on the two complementary products (which the individual firms were presumably not able to do pre-merger), and this will boost the merged entity's investments compared to the pre-merger scenario. This will necessarily discourage investment by their potential rivals, and whether or not this is harmful to consumers will depend on how large the investment post-merger will be in the aggregate (meaning the sum of the increased investment by the merged entity and the reduced investment by the outsiders), and how much of the benefits created by these additional benefits will be passed on to consumers.

(747) The Notifying Party observes that that model does not take into account the element of technology licensing, and that it ignores the fact that there are already rival suppliers of both LTE baseband chipsets and of NFC/SE chips active in the market. The presence of those alternative suppliers ensures that any entrant into either of the two component markets would be able to realise sales, because the necessary complementary component would be readily available, even if the merging party were to engage in pure bundling of its own products. Moreover, as far as the next generation baseband technology is concerned, the Notifying Party submits that numerous incumbents active in the LTE baseband chipsets market have already made major investments into these new technologies, so that there could be no deterrence effect on these already sunk investments from any pure bundling by the merging party.

(748) The Commission considers that the model confirms that pure bundling is an inherently costly strategy, because the firm engaging in this strategy has to sacrifice some profitable sales of standalone components it would otherwise have made, and this can only be profitable if bundling allows the bundling firm to compensate for these lost profits in some other way, for instance by increasing its chances to foreclose competitors. The Commission furthermore notes that the submitted model

639 Technical integration is one way in which a firm may render a strategy of pure bundling "credible", in the sense that it prevents itself from unbundling the two products later on, even if it were in the merged entity's interest to do so.

640 See "RBB response to further CRA submissions" of 06 September 2017. [DOC ID: 2559 ].

142

does not incorporate that some mobile OEMs have the possibility to source BCs or NFC/SE products internally, in which case bundling can be expected to be even less effective in deterring rivals from innovating. This would, in the first place, further reduce the incentives of the merged entity to engage in pure bundling. Furthermore, the profitability of bundling depends critically on the fraction of the value of its innovation which any single innovator can appropriate absent bundling, which is difficult to measure and to assess in practice.

(749) The submissions do not provide arguments to substantiate the key conditions that must hold for pure bundling to be a profitable exclusionary strategy, namely: (2) the third parties have not yet developed a product (and they may fail to do so even if they invest considerable resources into such product development), while the merged entity has its products already fully developed, so that its presence in the product market is to be taken for granted, and can invest at most in a marginal quality improvement; based on the available evidence the Commission considers that any asymmetry between the merged entity and third party rivals is less pronounced than assumed in the model. There appear to be alternative suppliers (including in-house supply for some OEMs) that could supply both BCs and NFC/SE chips. Moreover, the merged entity must be able to credibly commit to tying, which means designing its product in a way that makes it impossible to unbundle the product later on. In this respect, it could be possible for the merged entity, even in case of technical integration of the relevant components, to feature its integrated product in a way that would allow customers to use it in connection with a third-party component.

(750) Therefore, the Commission concludes that the economic submissions received do not suggest that post-Transaction, the merged entity would have the incentive to engage in pure bundling of LTE baseband chipsets, NFC and SE chips.

(751) In the in-depth investigation, the Commission also reviewed the Parties’ internal documents. As illustrated by the internal documents discussed in recitals (608), (609), (611), (612) and (617) above, the Parties in several instances discussed the integration of NXP’s products with the Qualcomm LTE baseband chipset, in particular the SE chip on the baseband chipset.

(752) From the content of those discussions, it appears that, while planning component integration, the Parties aimed to maintain open a parallel sales channel for a standalone solution, as discrete products would still play a role for device OEM customers. Those discussions therefore suggest that the merged would be unlikely to have the incentive to engage in either conducts of tying or pure bundling.

(753) In that regard, Qualcomm’s […] raised the opportunity of SE integration and selling the bundled product including MIFARE, but added to […] (emphasis added) .

(754) The same consideration for maintaining a sale of discrete components is reflected in NXP’s document […] which outlined the main areas of potential synergy between the Parties’ businesses. With respect to NFC and SE and LTE baseband chipsets, the document mentioned the plan to […] (emphasis added) .

641 NXP internal document, email from Qualcomm’s […] to NXP’s […], “[…]”, dated 8 February 2017, [DOC ID 1458-35882] [[…]].

642 NXP internal document, email from […] to Qualcomm’s […], “[…]”, dated 22 February 2017, [DOC ID 1452-11924] [[…]].

643 NXP internal document, “[…]”, pdf document attached to the email from […] to Qualcomm’s […], “[…]”, dated 22 February 2017, [DOC ID 1452-11924] [[…]].

143

(755) From other internal discussions, it also appears that Qualcomm was wary of the consequences of no longer selling components standalone. For instance, in a discussion over post-merger strategies on offering an integrated product consisting of the Qualcomm LTE baseband, the SE chip and MIFARE, Qualcomm’s […] noted that […]. This comment further highlights that, whereas Qualcomm would likely integrate, it would not be willing to entirely forego the sale of discrete standalone products.

645 (756) Similarly, Qualcomm’s […] commented that […] (emphasis added) .

(757) Internally, NXP also commented that […] among others because […] (emphasis added) .

(758) Accordingly, the Commission notes that, as already explained in Section 7.4.2.2 B) 2. above, the Parties’ internal documents suggest that Qualcomm and NXP would have the incentive to engage in mixed bundling conduct (including with the technical integration of the SE on the LTE baseband chipset), but not of pure bundling nor of commercial or technical tying, given that the Parties acknowledged that discrete components would still play a role and a separate sales channel for standalone components would be maintained.

(759) Furthermore, the Commission notes that following its acquisitions of other chipset manufacturers, Qualcomm did not engage into tying or pure bundling and continued to offer standalone products. For instance, following the acquisition of chipset manufacturer Atheros in 2011, Qualcomm continued to offer standalone WI-FI and Bluetooth chips .

(760) Therefore, based on the overall findings of the in-depth investigation, notwithstanding the feedback from market participants, the Commission concludes that the merged entity would be unlikely to have an incentive to engage in conducts of tying or pure bundling, given the lack of economic evidence and the findings in the Parties’ internal documents .

3. Likely effects on competition

(761) Notwithstanding the finding that post-Transaction, it is unlikely that the merged entity would have the incentive to engage in conducts of tying or pure bundling, the Commission notes that in any event such conduct, if implemented, would be unlikely to lead to significant foreclosure effects.

(762) In this Decision, in light of continuous technological improvements, the Commission considers that BC and NFC/SE suppliers need to continuously invest in R&D in order to be able to offer competitive products. Therefore, the Commission has in particular assessed whether bundling/tying would entail a significant risk that

644 Qualcomm internal document, internal email exchange entitled “[…]”, email by […], dated 11 October 2016, [DOC ID 2476-882].

645 Qualcomm internal document, internal email exchange entitled “[…]”, email by […], dated 5 November 2016, [DOC ID 2364-72772].

646 NXP internal document, internal email exchange, email sent by […] on 20 February 2017, “[…]”, [DOC ID 1456-51159].

647 RBB Economics paper of 18 August 2017, "Assessment of Qualcomm's selling practices following the acquisition of Atheros", pp 9-10, explaining that Qualcomm sells WCN connectivity chips (semi-integrated with the Qualcomm baseband chipset) and QCA connectivity chips, sold standalone.

648 For the sake of clarity, antitrust rules, in particular article 102 TFEU will continue to apply to the merged entity after the closing of the Transaction, regardless of the outcome of the present assessment under the Merger Regulation.

144

competing BC or NFC/SE producers would be hampered in their incentives or ability to invest in R&D and eventually could be foreclosed of the respective markets.

(763) In this context, the Commission notes that a number of market characteristics already set out in the context of mixed bundling suggest that foreclosing competitors is difficult in the concerned markets. The Commission considers that the findings made in Section 7.4.2.2 B) 3., in relation to mixed bundling of LTE baseband chipsets, NFC chips and SE chips, apply, mutatis mutandis, also in the context of a tying or pure bundling scenario by the merged entity.

(764) As mentioned in Section 7.4.2.2 B) 3. , device OEMs have not always decided to purchase a bundled product from Qualcomm, rather favouring a standalone product, and have obtained from Qualcomm the necessary technical support for mix-and-matching Qualcomm's baseband chipsets with other products, such as the Wi-Fi and Bluetooth chipsets of other suppliers. In the context of pure bundling, where the merged entity would cease to provide its components standalone, device OEMs would still have an interest to buy standalone components, rather than purchasing the merged entity’s bundle.

(765) Moreover, as recalled in recitals (651) to (658), customer behaviour appears to be inconsistent with foreclosure effects materialising as a result of a mixed bundling strategy of LTE baseband chipsets, NFC, and SE chips and standalone competitors can also have recourse to “mix and match” solutions.

(766) Finally, as recalled in recitals (536) and (665), based on the data submitted by the Notifying Party, the growth in the near future of NFC and SE on mobile devices with a non-Qualcomm baseband chipset will also contribute to standalone competitors of the merged entity having viable options to remain active.

(767) Therefore, the Commission considers that, even if the merged entity were to have the incentive to engage in tying or pure bundling, there is no sufficient evidence in the Commission’s file to conclude that such conducts would likely lead to significant foreclosure effects .

7.4.4. Degradation of interoperability

(768) Concerns were raised during the phase I market investigation by several respondents, both competitors and customers, that the merged entity would have the ability and the incentive of degrading the interoperability of Qualcomm's LTE baseband chipsets and of NXP's NFC and SE chips with rival suppliers' standalone components. The effect of such a strategy would be that device OEMs' customers would prefer the merged entity's product over those of rival suppliers. These suppliers would be eventually foreclosed from the market.

(769) As a general principle, for interoperability to work seamlessly, support by both sides is required both at the physical (hardware) level and at the protocol stack level (that is to say, the software implementation of the communication protocols). Therefore, interoperability support is require from both the baseband chipset supplier, as well as from the NFC and SE chips supplier.

649 For the sake of clarity, antitrust rules, in particular article 102 TFEU will continue to apply to the merged entity after the closing of the Transaction, regardless of the outcome of the present assessment under the Merger Regulation.

650 See response to question 56 of Q1 - Questionnaire to device OEMs; responses to question 49 of Q2 – Questionnaire to baseband chipsets competitors; and responses to question 56 of Q3 – Questionnaire to NFC competitors.

145

(770) Interoperability between the NFC controller (referred to as CLF (contactless frontend) in Figure 10 below) and the embedded SE is ensured via the Single Wire Protocol (SWP) interface, and between the (integrated) baseband chipset and the NFC/SE solution via the IC interface and the Serial Peripheral Interface (SPI).

Figure 10 - Interoperability between the NFC and SE chips and BC/AP

(771) Some smartphones also have the Trusted Execution Environment (TEE) which is a secure execution area of the application processor, separate from the Rich OS (the operating system of the smartphone for example Android). One respondent to the phase II market investigation argued that TEE represents a viable alternative to SE and that post-Transaction the merged entity would have the incentive to foreclose TEE providers.

7.4.4.1. The Notifying Party's views

(772) The Notifying Party submits that the merged entity would not degrade interoperability between its baseband chipsets and the NFC and SE chips of other suppliers, or vice versa.

(773) First, the Notifying Party explains that the NFC/SE chip is linked to the baseband chip through generic buses (I²C & SPI). These buses are based on worldwide standards, and as such, cannot be degraded for rival NFC and SE chips. This link is just a transport layer which has no effect on performance.

(774) Second, the Notifying Party claims that there are no past instances where it degraded interoperability for competing products.

(775) Third, the Notifying Party submits that degrading interoperability would be immediately visible to customers. This would be untenable commercially. Customers have sufficient countervailing buyer power to prevent the combined company from attempting any foreclosure strategy. As explained, customers value cross-technology interoperability and the ability to mix-and-match solutions from different suppliers so as to better meet their technical requirements and differentiate their products.

651 2 The Inter-integrated Circuit (IC) Protocol is a protocol intended to allow multiple “slave” digital integrated circuits (“chips”) to communicate with one or more “master” chips. Like the Serial Peripheral Interface (SPI), it is only intended for short distance communications within a single device.

652 See submission to the Commission by […], of 23 October 2017. [DOC ID: 3209].

653 Having examined the complaint, the Commission considers that the theories of harm put forward are not specific to the Transaction. Pre-merger, NXP does not provide access to its NFC chips to TEE providers and Qualcomm does allow access to its chipsets other TEE providers. The situation would therefore remain unchanged with the Transaction, should the merged entity refuse access to its NFC chips in the future.

146

7.4.4.2. Commission's assessment

A) Ability to engage in degradation of interoperability

(776) The Commission first notes that, based on the results of the phase I and phase II market investigations, while the SPI and SWP are standardised interfaces, this only applies to the physical (hardware) layer (how to get the data from one place to the other).

(777) Past this layer, on the transport layer, the technology is proprietary (for example, the formatting of the data is not standard). The commands that are sent through the SPI interface are not standardised (but proprietary) and need to be exchanged between the suppliers of the processor and the NFC/SE for the components to communicate with each other. The availability of the commands to the suppliers of the NFC/SE components is a pre-condition for the components to work together. Also, SE/OS products are offered together with a driver (software code) that needs to be implemented on the baseband chipset or the application processor and the SE/OS. The command set provided by the eSE/OS to the baseband or application processor then needs to be implemented on the baseband or application processor side. This requires interaction between the two parties: either the mobile device OEM using a specific baseband or application processor or the baseband/application processor vendor and the eSE/OS supplier. This interaction usually takes place during the design-in cycle of a mobile device. Interoperability between the components is ensured by exchanging command and data structure information as well as by joint bring-up or test sessions (collaboration between the suppliers).

(778) Therefore interoperability information and support is necessary from both the baseband chipset manufacturer's side, as well as from the NFC and SE chips manufacturer's side.

(779) Therefore, despite the Notifying Party's claims that those interfaces are standardised, the Commission notes, based on the responses to the phase I and phase II market investigations, that past the transport layer, the technology is proprietary. It would be thus necessary for the merged entity to provide interface related proprietary information and integration support for its baseband chipsets and NFC and SE chips which it sells standalone and which customer OEMs would like to mix-and-match with other suppliers' products.

(780) The Commission also considers that in the next couple of years when the merged entity would be expected to integrate NXP's SE chip on the baseband chipset, such integration process may result in some changes to the current interfaces between the baseband chipsets and the NFC and SE chips. Respondents to the phase II market investigation expect that the merged entity would reengineer those interfaces. Some respondents pointed out that there is no guarantee that the SPI and I²C interface would be maintained for the merged entity's integrated product. It is also possible that the SWP interface may have a different configuration. These respondents expect that for discrete mix-and-match solutions combining a third party NFC/SE solution with Qualcomm's integrated chipset, the existing difficulties in ensuring seamless communication and interoperability through the SPI may worsen.

654 See Infineon's response to RFI 29 of 5 July 2017, question 1.27, [DOC ID: 2955].

655 See minutes of the conference call with Infineon, of 11 September 2017. [DOC ID:2734 ].See also Minutes of the conference call with Gemalto, of 15 September 2017. [DOC ID:2835].

147

(781) Second, respondents to the phase II market investigation explained that the merged entity could degrade interoperability in more than one way. For example Infineon explained that: "Even without integration, Qualcomm would be able to degrade interoperability between the Qualcomm processor and third party NFC/SE as Qualcomm controls the interfaces and to the benefit of their solution both the processor and the NFC/SE. This could be done, for example, by ensuring that the Qualcomm processor performs transactions with a Qualcomm SE, i.e. responding to a "call" from the Qualcomm SE to the processor, at a frequency of every clock cycle but limiting that frequency of responding to a "call" from a third party SE to every other clock cycle, thus limiting the performance of mix-and-match solutions. This could be done by creating such software barriers on the processor side and could be done even if the necessary commands for interoperability are provided. This conduct can be detected only at a very late stage for example by measuring the answering speed to "calls" from a third party SE vs. "calls" from Qualcomm own SE. While a third party supplier would be able to detect such performance degradation, it is not able to prevent Qualcomm from doing so or to provide an alternative implementation without such performance degradation". Gemalto argued that: "There are several ways in which Qualcomm could degrade interoperability between their baseband processor chipsets and other NFC/SE solutions. The most extreme one would be for Qualcomm to integrate the NFC front end and the SE in a single baseband or application processor chip without providing for any interface to external NFC/SE solution, thus making it impossible for any NFC/SE solution provider to compete anymore on this market. However, there could also be more subtle ways to degrade interoperability: "The merged entity would prioritise improving the user experience for its integrated solution which would offer a better user experience (through e.g. improved timing and performance), and de-prioritise third party discrete mix-and-match solutions. Third party NFC/SE solutions could be further disadvantaged by Qualcomm further tightening the already very challenging timing and performance requirements imposed on embedded SE suppliers wishing to ensure communication with the Snapdragon via the SPI interface". Another chipset manufacturer indicated that degradation of interoperability could take the form of technical integration: "Qualcomm would make efforts to integrate NFC/SE more tightly with baseband chipset and claim it as total solution to smart phone OEMs".

(782) Third, the Commission considers degradation of interoperability could be the result not only of Qualcomm intentionally reengineering interfaces in such a manner as to degrade the performance of third party products, but also by not providing necessary information and support to ensure interoperability in the first place. Seamless interoperability between the merged entity's baseband chipsets/application processors and other suppliers' NFC and SE chips and vice versa would involve close cooperation between the merged entity and these third party suppliers. One respondent to the phase II market investigation explained how the merged entity would need to cooperate should a mobile device customer prefer a "mix-and-match" solution. Infineon explained that: "Qualcomm would need to make available to third party NFC/SE suppliers the necessary interface specifications and commands at an early stage of product development. Towards the end of the development stage, Qualcomm would also have to cooperate and provide support in the integration testing phase to ensure that the products communicate seamlessly (e.g., through ad-hoc meetings). Some large mobile device OEMs may be able to ensure interoperability testing by themselves, however the technical risks, the risk of delays in the time to market as well as the additional costs would be high". Therefore, the Commission considers that the merged entity would also have the ability to hamper with the mix-and-match choices of mobile device OEMs if not only by refraining from providing the necessary interoperability information and commands to third party suppliers. Since such information and commands are proprietary to Qualcomm, the merged entity could simply refuse to provide such information and commands, thus discouraging the OEMs to mix-and-match and instead prefer the merged entity's integrated or bundled product.

656 See minutes of the conference call with Infineon, of 11 September 2017. [DOC ID: 2734]

657 See Gemalto's response to question 3.1 of RFI 25 of 4 July 2017. [DOC ID: 2152].

658 Minutes of the conference call with Gemalto, of 15 September 2017. [DOC ID:2835].

148

(783) Fourth, the Commission also notes that during the phase I market investigation not only third party suppliers, but also some mobile device mobile OEM customers raised concerns that the merged entity would have the ability to engage in degradation of interoperability between its baseband chipsets and third party suppliers' NFC and SE chips and vice versa. One mobile device OEM customer argued that the merged entity could change interoperability between the software on the application processor, which manages SE chips and reads storage data in SE chips, so as to hamper the use of third party SE chips.

(784) The Commission thus considers, based on the results of the phase I and phase II market investigations, that the merged entity would have the ability to degrade interoperability between Qualcomm's baseband chipsets and other suppliers' NFC and SE chips or between other suppliers' baseband chipsets and NXP's NFC and SE chips.

B) Incentive to engage in degradation of interoperability

(785) For the reasons indicated in the following recitals, the Commission also considers that the merged entity would have the incentive to engage in interoperability degradation with a view to redirecting demand for third party standalone baseband chipsets and NFC/SE products to its own LTE baseband/NFC/SE bundle.

(786) First, respondents to the phase I and phase II market investigations, both competitors and customers pointed out that the merged entity would have incentive to engage in such a strategy.

(787) Thus, several NFC and SE chips suppliers argued that the merged entity would have such incentive. G&D submitted that: "There could be a motivation to degrade interoperability to impede or slow down integration components from other suppliers resulting in a competitive advantage as other suppliers may miss design-in Windows. A complete abundance of interoperability is not expected as OEMs would urge for alternatives. Notwithstanding, there could be undocumented optimizations which would make Qualcomm's solution superior while maintaining compatibility for others."

660 See minutes of the conference call with Infineon, of 11 September 2017. [DOC ID: 2734].

661 See responses to question 56 of Q1 - Questionnaire to device OEMs.

662 See submission by a mobile device OEM, of 2 October 2017, page 11. [DOC ID: 3138].

663 See response to question 56 of Q1 - Questionnaire to device OEMs; responses to question 49 of Q2 – Questionnaire to baseband chipsets competitors; and responses to question 56 of Q3 – Questionnaire to NFC competitors. See also RFIs 25 and 29 to Gemalto [DOC ID: 2152] and Infineon [DOC ID: 2955], as well as minutes of the conference calls with Gemalto, of 15 September 2017 [DOC ID: 2835] and Infineon, of 11 September 2017 [DOC ID:2734].

149

(788) Infineon argued that the merged entity could degrade interoperability so that such strategy would be to the benefit of the merged entity's baseband chipset and own NFC/SE solution.

(789) Among the device mobile OEM customers, […] argued that: […] .

(790) Second, the phase I and phase II market investigations provided some indications that do not support the Notifying Party's claims that it had never tried degrading interoperability or refuse to provide necessary support and information to other integrated third party products with its baseband chipsets and application processors in the past. Some suppliers did point out from their past experience with Qualcomm that it does not typically support OEMs in changing the reference design of its chips to support third party components. […] argued that […] .

(791) As explained in recital (57), the baseband chipset is one of the essential components of smartphones, as it ensures the connection of mobile devices to mobile telecommunication. Therefore it represents one of the most important choices that the OEM has to make when designing a new smartphone. This is also reflected by the price of the baseband chipset which is much more significant than that of the NFC or SE element. As a result, customers who choose Qualcomm's baseband chipset would be less inclined to purchase the NFC/SE solution of another supplier in a situation where the merged entity would degrade interoperability or not ensure the same level of interoperability between its baseband chipsets and the NFC/SE solutions of third party suppliers. Moreover, the importance of the baseband chipset, relative to the NFC/SE chip, makes it unlikely that a customer would entirely switch away from the merged entity's product, and in particular would start buying third party baseband chipsets, just to be able to combine it with its preferred third party NFC/SE chips.

(792) Fourth, as explained in recital (777), ensuring interoperability also means that the merged entity would need to provide the necessary information, commands support and cooperation to third party suppliers of standalone products. To the extent that

664 See Responses to Questionnaire Q3 to NFC Competitors, question 56.3.

665 See Minutes of the conference call with Gemalto, of 15 September 2017 [DOC ID 2835].

666 See Minutes of the conference call with Infineon, of 11 September 2017. [DOC ID: 2734].

150

providing such information and support to third party suppliers is costly to the merged entity, the merged entity would have less of an incentive to bear these costs. This is because as a consequence of the Transaction, the merged entity would also have in-house production of NFC/SE chips, and therefore internalise any profit gains from diverting BC customers from third party suppliers to its own NFC/SE products. The same reasoning also holds for diverting NFC/SE customers from buying third party BCs to buying BCs of the merged entity. Therefore, the merged entity would benefit post-Transaction, if, as a consequence of poor interoperability with third party suppliers, BC customers would also buy NFC/SE chips from the merged entity (and similarly for NFC/SE customers). The Commission therefore considers that the merged entity would find it less profitable to invest in supporting third parties' products to successfully interact with its LTE baseband chipsets and its NFC/SE chips respectively, compared to the pre-merger situation, where the Notifying Party did not have any in-house production of NFC/SE chips.

(793) For these reasons, the Commission takes the view that the merged entity would have the incentive to degrade the interoperability or not ensure the same level of interoperability between its baseband chipsets and the NFC/SE chips of third party suppliers and vice versa.

C) Likely effects on competition

(794) Based on the results of the phase I and phase II market investigations, the Commission considers that the merged entity's strategy of degrading interoperability would likely compound the foreclosure effects of the merged entity's strategy of increasing royalties for MIFARE or ceasing the licensing of MIFARE altogether to competitors, against the backdrop of mixed bundling of LTE baseband chipsets and NFC/SE chips, as described in Section 7.4.2 above. The refusal to provide the necessary information or commands to third party suppliers of standalone NFC/SE chips or baseband chipsets would have a similar effect on third party suppliers as degradation of interoperability.

(795) First, the Commission notes that during the phase I market investigation respondents among both competitors and customers emphasised that a strategy of degrading interoperability would have a negative impact on the ability to compete of rival suppliers of baseband chipsets and NFC/SE solutions, eventually foreclosing them from the market.

(796) Intel states that "[d]egradation of Qualcomm's baseband chipsets with competitors' products such as NFC/SE chips, would likely harm the competitiveness of the remaining NFC/SE chip suppliers, which Intel understands account for a small share of the business".

672 (797) […] indicates that: […] .

(798) Samsung LSI points out that: "Given that Qualcomm would have a dominant position in multiple relevant markets, if Qualcomm did pursue a strategy to degrade interoperability with its own components, then such a strategy will have an exclusionary effect on SLSI and other competitors in the market for the other components even if Qualcomm did not in parallel pursue an integration strategy. That is because the various components (say AP and baseband or NFC and baseband) need to interoperate and communicate to some degree. If Qualcomm’s baseband would not work and communicate with an SLSI NFC, then customers who use Qualcomm basebands could not use SLSI NFC".

670 See responses to question 50 of Q2 – Questionnaire Q2 to baseband chipsets competitors and Responses to question 56.4 of Q3 – Questionnaire to NFC competitors.

671 See Intel's response to question 50.1. of Q2 – Questionnaire to baseband chipsets competitors, [DOC ID: 767].

672 See […] response to question 56.4 of Q1 - Questionnaire to device OEMs, [DOC ID: 1009].

151

(804) Therefore, the Commission considers that such a strategy on the part of the merged entity to degrade interoperability of third party baseband chipsets and NFC and SE chips would compound the foreclosure effects stemming from an increase of MIFARE royalties (or a denial to license MIFARE), on top of a conduct of mixed bundling, as described in Section 7.4.2 above.

7.4.5. Conclusion on conglomerate effects in relation to Qualcomm's baseband chipsets and NXP's NFC and SE chips

(805) Based on its assessment in section 7.4.2, the Commission therefore takes the view that the merged entity's conduct of increasing royalties for MIFARE or ceasing the licensing of MIFARE altogether to competitors, applied in addition to mixed bundling of LTE baseband chipsets, NFC and SE chips, would have the likely effect of foreclosing competitors of baseband chipsets and NFC and SE chips.

(806) Furthermore, based on its assessment in section 7.4.4, the Commission takes the view that the merged entity's strategy of degrading interoperability would likely compound the anti-competitive effects of the merged entity's strategy of an increase of MIFARE royalties (or a denial to license MIFARE) with mixed bundling.

(807) In light of these conglomerate effects, the Commission considers that the Transaction as originally notified would give rise to a significant impediment of effective competition in relation to the markets for LTE baseband chipsets and NFC and SE chips.

7.5. Conglomerate effects related to IP licensing of NFC technology

(808) The Parties hold significant IP rights in particular in NFC technology. In light of the complementary nature of the technology involved and the manner in which IP licenses are negotiated with potential licensees, the combination of Qualcomm's and NXP's NFC patent portfolios could, absent suitable commitments, raises competition concerns relating to a disproportionate increase in bargaining power and the negotiated royalties.

(809) In the following Sections, the Commission will provide preliminary remarks on the patent licensing and on the Parties' pre-merger licensing practices before summarising the results of the phase I and phase II market investigations and the Notifying Party's view. The last Section provides the Commission's assessment. The Commission will analyse whether as a consequence of the Transaction, the merged

679 See Minutes of the conference call with Infineon, of 11 September 2017. [DOC ID: 2734].

680 NXP also holds […] SE-related patents and […] SE OS-related patents. NXP does not license its SE patents and monetizes its SE technology solely through the sale of chips. Qualcomm, to the best of its knowledge, does not hold any patents pertaining specifically to the SE of an NFC/SE chips solution (Notifying Party's Response to RFI 34 of 14 July 2017, question 8 [DOC ID 2333]). Therefore, the Transaction will not combine patents on SE technology. Moreover, the Commission notes that that NXP's SE patents are not a source of revenue in and out of themselves as the value of SE patents is reflected in NXP's sales of SE chips and its patents are not subject to licensing agreements (see NXP's Response to RFI 35 of 18 July 2017, paragraph 20 [DOC ID 2009]). The following theory of harm therefore does not extend to SE patents.

153

(811) Implementers who develop, produce or sell products may obtain licences from patent holders that permit the implementer to use a patented invention. In return for granting a license, patent holders typically charge royalties. The Commission's present concerns relate to the level of royalties that may be set by the merged entity for certain patents. It therefore requires explaining how royalties are generally negotiated and possible practices which, in the present case, could lead to conglomerate effects.

(812) In case no mutually acceptable licensing terms can be found, IP holders can remedy IP infringements by initiating court proceedings. At the request of the IP right holder, a court may order that appropriate measures be taken with regard to goods found to be infringing an IP right. Those corrective measures include injunctions banning the infringing products from the market. Moreover, courts can order damages against an infringer. In the context of a patent infringement suit, the court must also determine whether the relevant patents were actually infringed. Moreover, implementers involved in such litigation usually challenge the validity of patents allegedly infringed. Litigation costs as well as the resources needed to carry out infringement suits can be considerable.

(813) Licensing negotiations are usually done in the "shadow of litigation", that is, the licensing terms typically depend on the chances that at least one patent under negotiation will be found to be valid and infringed, as well as remedies imposed by a court. The chances for an infringement to be found in turn are a function of the strength and the number of infringed patents. For example, if a patent holder owns few contestable patents that might be found invalid, the chances for at least one patent to be found infringed may be small. In case of an infringement, a patent holder can be entitled to collect damages. In addition, depending on the circumstances, courts may issue injunctions to prevent further infringement. In case of an injunction, infringers should cease to sell infringing products and remove them from the market, thus losing any related profit.

(814) In the case of standard-essential patents (SEPs), once a standard incorporating proprietary technology is adopted, the potential exists for opportunistic patent holders to insist on patent licensing terms that capture not just the value of the underlying technology, but also the value of standardisation itself. To address this “hold-up” risk, the IP policies of standard setting organisations (SSOs) generally require patent holders to disclose their patents and commit to license SEPs on fair, reasonable, and non-discriminatory (“FRAND”) terms. SEP-holders' ability to

681 Case 187/80 Merck & Co Inc. v Stephar BV and Petrus Stephanus Exler [1981] ECR 2063, paragraph 4.

682 Case AT.39985 - Motorola - Enforcement of GPRS standard essential patents, Section 4.1.

683 See also the Commission Guidelines on the applicability of Article 101 of the Treaty on the Functioning of the European Union to horizontal co-operation agreements (the ''Horizontal Cooperation Guidelines''), paragraphs 285-286, OJ C 11, 14.1.2011, p.1. SSOs sometimes refer to "reasonable and non-discriminatory" terms. The present decision will refer to FRAND or (F)RAND terms.

154

(815) It follows that several factors affect the bargaining position of the parties in licensing negotiations and the ensuing licensing terms. These include the value of the infringing products or, in case an infringement can be avoided by redesigning or removing certain features, the incremental value that is added by the infringed technology. Furthermore, the number and strength of infringed patents is relevant, as these impact the likelihood of an infringement being found in the first place.

(816) Moreover, litigation-related factors may have an impact on the negotiated licensing terms. Specifically, an infringer may doubt in the first place that an IP holder will initiate litigation, for example due to the low chances of being awarded any remedy, litigation costs, lack of resources to (successfully) carry out such litigation, or likely significant negative consequences, such as counter-suits or loss of valuable other business. Therefore, the relative litigation strength of the two parties can influence their bargaining position and the agreed royalty level.

(817) Therefore, for the purposes of this Decision, the Commission considers that royalty levels set in patent licensing negotiations depend not only on the strength and the number of patents, but also on additional factors affecting the credibility of threats to initiate patent litigation.

(818) Against this background, the Commission examined whether the Transaction could give rise to conglomerate effects resulting from the combination of the Parties' respective NFC patents. In this respect, it should be noted that NXP does not grant NFC licenses on a patent-by-patent basis, but generally licenses third parties to a patent portfolio. From the perspective of prospective licensees, such licenses bundle together largely complementary patents necessary to implement NFC technology without infringing NXP's patents.

(819) By combining the Parties' respective NFC portfolios, the Transaction would enable the merged entity to offer licenses covering a broader portfolio of NFC patents. The mere combination of patents, even for licensing purposes, does not in itself raise competition concerns in principle; nor does a patent holder's use of factors contributing to its credibility in licensing negotiations. Stated differently, the mere combination of two previously separate patent portfolios, depending on the circumstances, may lead to an increase in bargaining power and royalties to the benefit of the acquiring patent holder, in the sense that the latter would largely equal the sum of the relevant patent holders' respective royalties pre-merger. Such change would not induce competition concerns.

687 This is without prejudice of the assessment of certain forms of portfolio under Article 102 TFEU.

155

(820) However, the Commission is concerned that, in the particular circumstances of this Decision, bundling the Parties' NFC patents as a result of the Transaction would enable the merged entity to profitably increase royalties above levels which the Parties could have charged absent the Transaction for the same patents.

(821) Whereas the Commission recognises that the combination of patent portfolios, depending on the circumstances, can have pro-competitive effects, the Notifying Party has not advanced any substantiated case-specific arguments relating to such potential efficiencies.