Running Target: Next-Level US Tech Controls on China
With the US midterm elections around the corner and the 2024 presidential race set to unfold after that, the White House is preparing the next phase of its China strategy, with a focus on preserving US technological leadership. The recently passed CHIPS and Science Act, which aims to revitalize US semiconductor manufacturing, forms the offensive prong of the Biden administration’s strategy. Now for the defensive components. From list-based export controls on specific technologies to the buildout of outbound investment screening, the White House is putting together a playbook with important implications for a range of actors—including US industry and America’s allies abroad. Among the key questions: Where should the US draw the line when defining technologies as “leading edge”? How far should it go in using “long arm” provisions that target certain Chinese entities? And what role should human rights, supply chain resilience, and data privacy concerns play in technology restrictions under a broader national security umbrella?
Unpacking the White House Agenda on Tech Controls
The summer was full of clues about the White House’s fast-evolving agenda on tech controls:
- August 9: Biden signed the CHIPS and Science Act into law, allocating $39 billion in subsidies for US semiconductor manufacturing (including $2 billion for legacy chips) and an additional $11 billion for boosting R&D.
- August 12: The Commerce Department’s Bureau of Industry and Security (BIS) issued an interim final ruleon Wassenaar-level multilateral controls, including for Electronic Computer-Aided Design (ECAD) software designed to develop integrated circuits with Gate-All-Around Field-Effect Transistor (GAAFET) structures—innovative 3D transistors that support node sizes at or below five nanometers (nm).
- August 31: US technology firm NVIDIAdisclosed in an SEC filing that, as of August 26th, it was being restricted from selling its most advanced AI-powered Graphics Processing Units (GPUs)—A100 and H100 chips—to China. US semiconductor company AMD confirmedthat it had alerted Chinese suppliers of the new licensing requirement. A Commerce Department spokesperson told Reuters that a “comprehensive approach” was being taken with regard to “technologies, end-uses, and end-users” in order to protect US national security and foreign policy interests.
- September 1: NVIDIA disclosed in an SEC filing that as of August 31st, it had received licenses to export newly restricted A100 chips in support of US customers through March 1, 2023 and A100 and H100 chips through its Hong Kong facility through Sept. 1, 2023, giving the company time to adapt to the new restrictions.
- September 2: Semafor reported that the White House was planning to move ahead with an executive order on outbound investment screening ahead of the midterm elections but was still debating whether to go beyond a notification requirement to include blocking authority.
- September 11: Reuters reported, citing unnamed sources, that the Commerce Department intended to formalize new requirements to restrict a) AMD and NVIDIA from selling advanced AI-related GPUs and b) KLA, Lam Research, and Applied Materials from selling semiconductor manufacturing equipment for sub-14nm logic chip production. The restrictions would apply to China and Russia.
- September 15: The White House issued an executive order providing presidential guidance on the risks CFIUS should consider when reviewing transactions, with a special focus on critical technologies, supply chain security, investment patterns, cybersecurity, and personal data protection.
- September 16: US National Security Advisor Jake Sullivan delivered a speech at the Special Competitive Studies Project Global Emerging Technologies Summit, drawing attention to “force-multiplying” technologies and asserting that the US needs to go well beyond the objective of staying “a couple of generations ahead” in strategic technology development.
These developments show that a more coherent US strategy on technology controls is taking shape, but one which risks verging into more extreme regulatory territory. A White House shift toward more list-based technology controls—against the backdrop of rising US frustration over the readiness of key allies to restrict tech exports and domestic political incentives to out-maneuver the Republicans on China policy—is likely to give rise to more trade volatility, including in global semiconductor supply chains.
Narrowing tech priorities
An important policy debate in Washington in recent years has been over how best to design export controls, both in terms of technology subsets and thresholds for defining “leading edge” technologies. The BIS has fended off repeated demands by lawmakers to issue a master “emerging and foundational technology list”. Meanwhile, the White House Critical and Emerging Technology List remains alarmingly broad. However, the executive order issued by the White House offering guidance to CFIUS, Sullivan’s recent remarks, and the leaks on the White House’s upcoming outbound investment screening order point to a narrowing priority list for critical technologies that includes: advanced semiconductors, artificial intelligence, and quantum computing (large-capacity batteries and biomanufacturing are also growing areas of focus.) These technologies will likely be the focus of new export controls and investment screening.
Broadening beyond entity-focused technology controls
The recently disclosed BIS restrictions on AMD and NVIDIA AI computing chips confirm a new focus on a smaller list of critical (or, to use Sullivan’s terminology, “force-multiplying” technologies.) Although NVIDIA was granted a license to fulfill pending orders for US customers, the intent was clear: Commerce is focusing on a number of cutting-edge technologies that it wants to keep out of China’s hands.
This could mean a shift in emphasis away from entity listings (an exasperating cat-and-mouse game of BIS chasing down spinoffs in labyrinthine party-state business ecosystems) to broader technology-based controls based on potential military end-use and applied against countries of concern, such as China and Russia. Ad hoc entity listings will remain a tempting shortcut for policymakers to ratchet up sanctions and score political points, but a deeper foray into list-based technology controls can have a much broader impact on transactions related to dual-use technologies.
As the White House sharpens its focus on “force-multiplying” technologies, the next logical step will be for Commerce to better define military end-use and end-users in justifying list-based technology controls. This could impose a much heavier regulatory burden on companies to make a determination on end-use if their products fall into a military-civil fusion tech ecosystem.
Defining “leading-edge” and “chokepoints” for advanced chips
The US administration is also creeping closer to defining thresholds for what constitutes the “leading edge” in “advanced” semiconductors:
- Export controls that BIS is reportedly formalizing would target semiconductor manufacturing equipment (SME) “chokepoints” by restricting US suppliers (KLA Corp, Lam Research Corp, and Applied Materials) from exporting equipment capable of producing logic chips at or below 14 nanometers (nm).
- The BIS interim final rule on Wassenaar controls restricts “Electronic Computer-Aided Design (ECAD)” software specifically designed for producing chips with Gate-All Around Field-Effect Transistor (GAAFET) structures. This is an advanced 3D transistor architecture that supports the production of cutting edge chips at or below 5nm.
- The CHIPS and Science Act sets the bar for “legacy chips” at or above 28nm, for the purposes of restricting recipients of CHIPS funding from engaging in “significant transactions” that involve the “material expansion of semiconductor manufacturing capacity” in China.
The signs so far suggest the White House wants to freeze China’s capabilities in place at 14nm (chips that China’s SMIC is capable of producing) at a time when multilateral controls are geared toward more advanced technology thresholds (5nm and below). The White House is consciously taking a more rigid approach toward defining the leading edge. As US National Security Advisor Jake Sullivan stated in a September 16 speech:
“On export controls, we have to revisit the longstanding premise of maintaining ‘relative’ advantages over competitors in certain key technologies. We previously maintained a ‘sliding scale’ approach that said we need to stay only a couple of generations ahead. That is not the strategic environment we are in today. Given the foundational nature of certain technologies, such as advanced logic and memory chips, we must maintain as large of a lead as possible.”
It can be reasonably inferred from the Sullivan comments and recent reporting on emerging controls that the Biden administration intends to set a more stringent threshold at 14nm in order to prevent advances in China’s logic chip development instead of designing semiconductor export controls to keep pace with innovation. Such an approach is designed to give the United States “as large of a lead as possible,” but it comes with important side effects. For example, the US administration is also trying to target perceived chokepoints in the semiconductor supply chain to hamstring China’s indigenous chip development. Semiconductor manufacturing equipment (SME) is a growing area of focus for this chokepoint strategy. But many SME components are not specially designed for sub-14nm production, and are used for producing both leading edge and legacy node chips. Wider ripple effects in semiconductor supply chains can therefore be expected if a blunt 14-nm standard is applied across export controls.
Lithography—a complex patterning process in chip fabrication—attracts the most attention in the SME chokepoint debate. Dutch firm ASML’s current generation of extreme ultraviolet (EUV) lithography machines uses a wavelength of 13.5nm. These machines are already restricted from export to China. But SMIC is reportedly producing at sub-10nm, using less-advanced deep ultraviolet (DUV) machines. There is a big difference between developing prototypes and producing leading-edge chips at scale. However, SMIC’s recent progress has nonetheless reinvigorated the US policy debate on whether the SME threshold should be lowered to restrict DUV as well. Such a restriction would stray further from “leading edge” controls.
The Biden administration has made a concerted effort to coordinate technology controls with its partners to avoid alienating allies and ceding market share to foreign competitors. To this end, it has created arrangements like the US-EU Trade and Technology Council, Indo-Pacific Economic Forum working groups, and the emerging CHIP4 alliance (including the United States, Japan, Taiwan, and South Korea) to try to harmonize technology development and controls.
Progress has been predictably slow, however. US tech partners are ready and willing to join the strategic conversation (partly out of fear of missing out) but would rather focus these forums on constructive themes (coordinating development) than on restrictions that risk provoking a backlash from Beijing or raising costs for businesses.
Slow progress on the multilateral front risks feeding US anxiety over technology competition with China, driving it toward broader controls, with an extraterritorial flavor. The areas we are watching as we assess the potential for US long-arm provisions include:
- Outbound investment screening provisions: A June congressional proposal for outbound investment screening led by Senators John Cornyn and Bob Casey contained broad language on covered activities carried out by any US person, foreign person, or their affiliates, including any entity “influenced” by a country of concern. The White House has instead been focused on plugging holes in existing export controls to prevent US private equity and venture capital from reaching entities of concern.
- The White House is likely to take a step-by-step approach to outbound investment screening, starting with a Treasury-run mandatory notification requirement, linked to the White House priority list of technologies (advanced semiconductors, AI, quantum computing, and possibly biomanufacturing and subsets of clean energy technologies like high-capacity batteries.) The White House may be tempted to broadly define the covered activities as it tries to prevent US companies from evading controls through foreign subsidiaries.
- BIS controls on “activities of US persons”: In addition to the Entity List Foreign Direct Product Rule, which was applied first to Huawei and then more broadly to Russia in cutting off the supply of US-origin technology, the US government has another powerful weapon that could come into play as it narrows its technology priorities. Under Section 744.6 of the Export Administration Regulations (EAR), BIS could impose comprehensive controls on the “activities of US persons” that support the development for specific dual-use technologies, military-intelligence end-use, or end-user. Given the pervasiveness of “US persons” linked to technology supply chains, from design to financing, the extraterritorial reach of such a measure would be extensive and could damage relations with allies producing advanced technologies.The all-encompassing national security umbrellaBuilding resilience in critical supply chains, preventing human rights abuses, protecting US citizens’ sensitive data, and bolstering US cybersecurity are all big objectives that overlap with the national security goal of preserving US technology leadership. In trying to address these concerns (and score political points along the way), it may be tempting for US policymakers to package policy objectives together under the national security umbrella when developing new controls. The US government may also end up targeting Chinese technology conglomerates on multiple grounds, straying further from the technical approach that usually informs export controls. The Biden administration has tried to differentiate itself from the haphazard approach of its predecessor in targeting Chinese entities. This approach was highly disruptive to businesses, prone to legal challenges, and polarizing to US partners. But the White House is still trying to link its broader objectives to its strategic tech agenda. For example, after launching an export controls and human rights agenda at the December 2021 Summit for Democracy, we expect the administration to roll out more on this theme ahead of the midterm elections. Going forward, this could mean a ratcheting up of long-arm sanctions on entities involved in surveillance and cyber intrusion technologies and focusing on Chinese tech conglomerates involved in developing “force-multiplying” technologies in the military-civil fusion ecosystem.
The White House is trying to balance its strategic goal of advancing US technology leadership with the political objective of getting out ahead of the Republican agenda on technology controls should Congress flip in November. As a result, we can expect a flurry of policy announcements this fall that underscore a “tough on China” approach.
The Biden administration would prefer to put the focus on export controls in its China strategy and avoid potentially provocative measures (such as the Taiwan Policy Act being developed in Congress) which could ratchet up tensions with Beijing. By quickening its pace on technology and related investment controls, the White House hopes to build an unassailable lead in the technology race with China and establish limits for US policymakers. This approach is based on several, still untested, assumptions:
- That China, already falling short of its technology self-sufficiency goals (especially in semiconductors) and undergoing a deep economic slowdown, will mostly absorb hard punches while struggling to innovate around controls. Beijing has mostly held back on retaliatory moves over export controls given its overwhelming external dependency on external tech inputs. However, China has extraterritorial tools—including anti-monopoly veto power on foreign tech mergers and anti-suit injunctions for standard essential patents—that could be applied more forcibly as the US tightens controls.
- That foreign partners will follow the US lead and align on controls rather than run afoul of US restrictions (or risk ceding the moral high ground on human rights). For more geopolitically exposed partners like Taiwan, the US may hold more sway. Japan will try its best to steer the US toward more pragmatic controls. But Europe may resist signing on to policies that are specifically aimed at Beijing on the grounds that space for engagement must be preserved. Plans by German Chancellor Olaf Scholz and French President Emmanuel Macron to travel to Beijing in the coming months to meet with Chinese President Xi Jinping attest to their desire to maintain dialogue.
- That non-US technology firms from like-minded countries won’t feel compelled to whittle down US content and develop non-US alternatives in critical tech supply chains, especially in the medium- to long-term.
- That new US regulatory tools, even if based on narrower technology priorities for now, will not end up enabling a more extreme approach to technology controls under a more hawkish US administration down the road.
The White House is moving beyond its mop-up of Trump-era policies to action on technology controls targeting China. In developing a more coherent framework for designing those controls, the Biden administration is taking a big step forward in defining the policy debate. But the pressure of election cycles can lead to policymaking that creates big spillover effects. As the US quickens its pace on tech controls, we are likely to see a ripple effect on US-China relations, existing US dialogues with like-minded technology partners, and on private industry in the months ahead.