Vol. 5 No. 13 - January 17, 2017 - Click here to access our library.
Given MOFCOM’s industrial policy-driven approach to merger review, and the companies’ semiconductor technology leadership and overall size, MOFCOM-related risks ultimately pose the biggest hurdle to the Qualcomm/NXP deal’s close. As the Chinese government actively seeks to reduce its technology gap with global technology leaders, MOFCOM, according to interviews with a number of semiconductor, CFIUS, and U.S.-China relations experts, is especially likely to target this deal, up to and including potentially demanding material divestitures as a condition of clearance.
A Capitol Forum review of the companies’ portfolios reveals that the combined firm will hold IP and design chips for several semiconductor sectors in which the Chinese government harbors significant ambitions, including power management and compound IC technologies. In addition, Qualcomm and NXP are market leaders in key innovation markets including automotive and IoT, where, given China’s stated ambitions to move up the global semiconductor value chain, MOFCOM may likewise seek asset sales.
Even if the parties are able to reach an agreement with MOFCOM to divest assets as a clearance condition, several of Qualcomm/NXP’s business segments, including wireless and mobile technologies, if divested to a Chinese acquirer, could raise CFIUS concerns due to their potential military applications and other national security issues. This is especially true given a number of recent developments which could shift CFIUS policy towards a more hawkish view on Chinese foreign direct investment, which in turn, could increase the likelihood of a tougher, retaliatory stance by MOFCOM.
In Depth: China’s Push in Semiconductor Technology and IP
MOFCOM likely to push for divestiture of attractive business units. Although MOFCOM has become somewhat more predictable in recent years, the agency is nonetheless a notoriously unpredictable enforcer, given its predilection for using merger review as a tool of state industrial policy. Considering the limited horizontal overlap between the parties, there is seemingly little antitrust justification for pushing for divestitures as a remedy for this deal— that is, antitrust within the U.S. or EU framework. MOFCOM’s Anti-Monopoly Law, on the other hand, calls for MOFCOM to consider industrial policy— particularly when considering deals where China’s key commodities and industries are at stake. Sources agree that it would be very much in character for MOFCOM to seek divestitures as a condition of clearing this merger, given the deal’s overall size, plus Qualcomm and NXP’s leading positions in China’s semiconductor industry.
The Qualcomm/NXP deal is likely to be of particular interest to MOFCOM, not only because of the parties’ extensive business activities in China, but also because the firms have leading edge technologies and IP in emerging sectors. “In silicon semiconductor technology (logic, mixed-signal, and both DRAM and 3D NAND memory), China seeks to reduce the technology gap with global technology leaders, such as Intel, Samsung, Qualcomm, and Micron,” says Dieter Ernst, a senior fellow at the East-West Center and an authority on innovation strategies and policies in the U.S. and China.
Rob Atkinson, founder and president of the Information Technology and Innovation Foundation (ITIF), and member of the U.S. State Department’s Advisory Committee on International Communications and Information Policy, notes that while private Chinese semiconductor firms tend to go after smaller companies, MOFCOM tends to target the biggest deals. “MOFCOM really has the incentive to go after the bigger companies,” says Atkinson. “That way, they get more bang for the buck, because bigger companies have more leeway in terms of valuable technologies they can give to the Chinese. Unless the next administration pushes harder, MOFCOM is not going to suddenly change. It’s seen a lot of valuable gains from this strategy.”
China’s newest semiconductor strategy covers wider range of products and value chain stages than previous initiatives. In March 2016, China revealed its 13th Five Year Plan (5YP), which focuses primarily on rebalancing China’s economy through more streamlined manufacturing and the development of advanced technologies. Among all high tech sectors, the Chinese government particularly stresses the importance of the information and communications technology (ICT) sector. China notes that ICT development will help improve efficiency in other areas of the economy, including industrial automation, smart manufacturing, and transportation.
In a recent East-West Center study, China’s Bold Strategy for Semiconductors, Ernst notes that previously, the developing Chinese semiconductor industry had been narrowly focused on ICs for mobile communication equipment and on semiconductor assembly, packaging and testing (APT). Now, on the other hand, China’s semiconductor industry strategy covers a much broader range of products and value chain stages, while at the same time increasing their depth and sophistication.
China’s current technology focus – memory, power management, and compound ICs. According to Ernst, “the [Chinese government’s] focus right now is on a big push in memory.” However, he says, Chinese firms are working to gradually reduce the technology gap in all segments, as opposed to a singular focus on memory chips. This view was echoed by another expert on China’s semiconductor strategy, who explained: “China’s focus is not limited to a single subsector. You see memory, you see power management, you see mobile applications – it’s a fairly broad mix.”
China’s other areas of interest in ICs include power management and compound IC technologies. Power management refers to the generation and control of regulated voltages which are required to operate electronic systems, while compound ICs refer to non-silicon semiconductor materials, such as Gallium arsenide (GaAs) and Gallium nitride (GaN), also referred to as wide-bandgap materials. According to Ernst, “in compound semiconductors, China seeks to leapfrog in LED [or light-emitting diode] technologies, based on GaN and SiC [or Silicon carbide] semiconductor materials, and hopes to narrow the technology gap in wide-bandgap power electronics.”
Neither Qualcomm nor NXP offer memory ICs, but NXP designs and manufactures both power management and compound ICs using its own IP and fabs. These assets appear especially likely to draw MOFCOM’s interest.
Global industry’s focus: automotive and Internet of Things (IoT). Beyond just closing the technology gap between itself and global leaders in the semiconductor value chain, China seeks to eventually forge ahead of global industry leaders and boost itself up in the global semiconductor value chain. By 2020, China hopes to be an industry leader in IC design (and China specifies design, over and above semiconductor APT or manufacturing) for key sectors including mobile devices, the IoT, and cloud computing. Similar goals have been highlighted in other major initiatives, including “informatization”— i.e., a strategy towards modernization and industrialization through IoT, cloud computing, big data, e-government, and e-commerce — and a new model of urban development— or “smart cities”— where ICT and IoT solutions merge to manage cities’ assets and services.
If China is ever to create innovative technologies, it will need stay at the leading edge of new technologies, while simultaneously working with technologies that other (especially U.S.) firms have already established. “As a latecomer, China understands that it needs to combine hybrid strategies that combine incremental innovations based on technologies that are one or two stages behind the frontier, while at the same trying to forge ahead in radical breakthrough innovations,” says Ernst.
In Depth: Most Likely MOFCOM-Directed Merger Conditions; CFIUS Implications
Qualcomm unlikely to agree to divest any of its core chipset or licensing business segments. Qualcomm has two primary business segments: chipset and licensing. Its chipset business involves chips used in voice and data communications, networking, and application processing. This segment includes LTE modems and Snapdragon processors, Qualcomm’s leading products. The chipset business accounted for 65% of Qualcomm’s 2016 revenues.
Qualcomm’s licensing business, at 33% of revenues, may be smaller in terms of revenues but is a more fundamental part of the company. Qualcomm collects license fees plus royalties, based upon the price of the end product (e.g., mobile phone) into which its chips get incorporated, rather than the price of the chips. Although Qualcomm has been facing pressure from antitrust regulators to change its royalty rate-setting practices—which would dramatically lower Qualcomm’s QTL earnings—and the company has recently faced numerous investigations and major fines in connection to its licensing business, it is unlikely that Qualcomm would agree to sell off any part of its wireless and mobile business, whether chipset or licensing.
Furthermore, even if Qualcomm were to agree, wireless and mobile technologies often have military applications, therefore raising the prospect of national security issues and CFIUS-related risks. In fact, when Qualcomm was first formed, it partnered with the U.S. military to develop secure mobile communications technologies (which just happened to be CDMA, the basis for 3G wireless).
Parties more likely agree to a divestiture of one (or more) of NXP’s business lines. NXP has four business segments:
1. Automotive (35% of revenue; 17.3% Y/Y increase): ADAS, infotainment, secure in-vehicle networking, electric vehicles;
2. Secure connected devices (SCD; 24% of revenue; 22.7% increase): processors, power management solutions, touch sensors, point-of-sale systems;
3. Secure interface and infrastructure (SI&I; 19% of revenue; 10% increase): networking processors, secure interface and system management products, high-performance RF power amplifiers; and
4. Secure identification solutions (SIS; 7% of revenue; 2.3% decline): security and privacy solutions for the banking, transportation, and e-government markets.
Of these segments, growth should continue to increase rapidly in NXP’s Automotive business and the SCD segment (which primarily delivers IoT solutions), assuming the post-merger firm is able to successfully ride the wave of growth in the automotive and IoT markets. With that in mind, Qualcomm will be hesitant to allow MOFCOM to chip away too much of these businesses; after all, NXP’s various automotive- and IoT-related chips are an integral part of the “complete system solutions” that the parties have touted as a key benefit of the combination, and forced divestiture of portions of these business could constitute a Company Material Adverse Effect.
Still, China could decide that the chance to extract desirable technologies, with this much leverage in a giant merger, is too rare an opportunity to waste—so it is possible that MOFCOM could require divestitures from NXP’s Automotive and/or SCD segments. To the extent that China targets business lines that are somewhat commoditized (e.g., lower-end processors), a sale to a Chinese firm could very well make its way through CFIUS review without issue, as did NXP’s sales of its RF Power and Standard Products businesses.
On the other hand, these segments contain a broad portfolio of products, including connectivity and security assets, IP, and application knowledge. If MOFCOM were to require the sale of higher tech, application-specific business lines, particularly with industrial, aerospace, or defense applications, CFIUS review could get into dicey questions about sensitive technologies and national security. This is even more likely when considering that NXP highlights the SCD segment’s applications in areas like mobile payments, healthcare, smart cities, and smart industrial.
Parties would most readily divest assets from NXP’s Secure Identification Solutions (SIS) and Secure Interface & Infrastructure (SI&S) segments. When considering revenues and growth, Qualcomm is likely most willing to divest NXP’s SIS segment. Revenues for the segment, which delivers security and privacy solutions to the banking, transportation, and e-government markets, have been falling since 4Q15. The parties may in fact welcome the sale of the SIS segment, given their tendencies to slough off business lines as they become less profitable and/or less connected to their core businesses (e.g., Qualcomm’s sale of its augmented-reality software platform Vuforia in 2015; NXP’s sale of its Standard Products division to JAC Capital, expected to close 1Q17).
For this segment, however, parties’ willingness aside, the question would be whether the SIS segment’s connection to banking, security, and privacy could potentially give rise to CFIUS concerns. NXP’s work with government agencies to create government-sponsored identification documents, health cards, and payment platforms under this segment could raise national security, privacy, and infrastructure concerns.
The parties would also probably be willing to divest an asset or two from the Secure Interface & Infrastructure segment, as SI&I had relatively slow growth (compared to automotive and SCD) and accounts for less than 20% of NXP’s total revenue. But it is important to note that at least two SI&I product lines have military applications, which automatically brings heightened CFIUS scrutiny.
First, NXP’s QorIQ processing platforms are complete system on chip (SoC) processors for networking applications across carrier, military, and industrial applications. Second, NXP’s high-performance RF (HPRF) power amplifier business has military applications, though NXP notes in its regulatory filings that the HPRF business targets mobile and industrial applications, “and to a lesser extent the military and aerospace markets.” According to NXP’s website, the “RF aerospace and defense portfolio offers high power solutions… that support a wide range of needs for defense and commercial applications,” including radar, communications, avionics, missile guidance, and “identification friend or foe” (IFF).
Commercial/commodity-based product lines could also be closely scrutinized. Atkinson notes that while there are certain types of semiconductor technologies that are “completely off the table,” such as specialized chips that are used by the military, “the U.S. government has been sending a signal that in some cases, you can't buy even the more commercially-based products either. Sure, there’s somewhat of a difference there, but I don’t think the difference is actually very meaningful,” says Atkinson. This is in part because many semiconductors are “dual-use” ICs, meaning that the same product can potentially be used in multiple applications. “What the Chinese want to get is defense capabilities—and you begin to get those capabilities by buying up commercial firms, and then using that to develop and get to other defense-related semiconductors. So even commercial technologies can be a stepping stone to get more capabilities for defense,” Atkinson explains.
Lastly, with respect to the potential for CFIUS-related issues for Qualcomm/NXP, it is important to note that CFIUS is only a stumbling block to the extent that the businesses and facilities that must be divested are located in the U.S. (at least under current CFIUS policy). So if MOFCOM were to require the divestiture of a NXP business totally unconnected to any of the nine NXP subsidiaries in the U.S., CFIUS review may not even be triggered.
MOFCOM more likely to require outright divestiture than JV or partnerships. According to sources, China is eager to have control of its own IP and technical standards. “China is really at the stage where they want to have control of what they call their ‘core technologies,’” says Richard P. Suttmeier, a professor emeritus at the University of Oregon and an expert on the role of technology in U.S.-China relations. “Licensing and JVs are good for getting access, but you don’t quite get all of the benefits of the technology that way. Semiconductor technology is something that is really not easily transferred, because of its complexity and expense. It’s something that really builds on years and years of accumulated experience, especially in terms of design, and I think that’s why the M&A route is appealing. There’s a very strong sense that China has to control the IP— or at least that’s where they want to be.”
Atkinson agrees that Chinese firms do not want to be dependent on foreign companies’ IP, particularly in the newest iteration of its semiconductor strategy. “First they had an indigenous strategy in 2006, where they were focusing on Chinese firms, not foreign companies; then came the exclusive semiconductor focus in 2009, where they got much more serious about the industry and were forming JVs and partnerships; then the purchasing tactic, around acquisitions, is even newer, in the last three years or so,” says Atkinson.
China’s ambitions of gaining and controlling core technologies and IP, rather than access through licensing and JVs, makes sense when considering China’s goal of rising in the industry value chain through IC design: in order to do so, it desperately needs domestic licensors of IC design-related IP. Other benefits of outright acquisition, over and above JVs and partnerships, include being able to avoid licensing fees and royalties (and conversely being able to charge other firms for these fees); entry into new markets and access to the non-Chinese market; and added R&D capabilities and manufacturing facilities.
Furthermore, Qualcomm and NXP have both already formed multiple JVs and partnerships with Chinese firms (in fact, Qualcomm has even invested directly in China’s leading IC fabrication company, SMIC), as both firms are eager to maintain and expand their positions in the Chinese market. Given that JVs and partnerships with Chinese firms are de rigueur for firms with a large market presence in China, such terms are much less compelling in terms of extracting valuable technologies through the merger review process.
Deal clearance could be complicated by a tougher MOFCOM stance, or tougher CFIUS reviews by the time deal is in late stage. At the recent CES 2017, NXP CEO Rick Clemmer addressed speculation that it could be more difficult for Qualcomm/NXP to get approval from MOFCOM if President-elect Trump’s hardline stance sparks a trade war. Clemmer apparently dismissed the possibility of China blocking the deal—while allowing that “if President Trump were to do something that would result in an increase of tension [with China], it could mean there might be a delay.”
As for CFIUS review, the next administration is likely to take a harder stance towards national security and major transactions in the semiconductor space—particularly deals with Chinese firms. Sources note that changes to CFIUS policy (which can be articulated by President-elect Trump) are more likely than a legislative overhaul—and certainly more likely to have implications in the near term.
One possibility is that CFIUS policy could shift towards a more expansive definition of national security, where openness of markets and trade reciprocity are included as part of CFIUS’s evaluation. Another possibility is a blanket ban on investment in the U.S. by Chinese SOEs, as recently urged to Congress by the U.S.-China Economic and Security Review Commission.
On January 6, the White House released a report entitled Ensuring Long-Term U.S. Leadership in Semiconductors, by the President’s Council of Advisors on Science and Technology (PCAST). The report puts forth another potential recommendation to strengthen CFIUS policy: “One way to respond would be to tie U.S. assessments of the national-security threats posed by particular technology exports, investments, and contracts to Chinese policy. (For example, if China pursues a policy of undermining cutting-edge, defense-critical U.S.-based companies by flooding markets using government support, that should later U.S. assessments of whether Chinese acquisitions of the capabilities required to do so are unacceptable).”
Between the PCAST report, other recent calls to strengthen CFIUS policy, and general concerns about the incoming administration and rising tensions with China, it seems that navigating CFIUS review—and MOFCOM review to a lesser (and related) extent—may become substantially more difficult in the upcoming months.