U.S. Semiconductor Export Controls on the PRC
Prospects for Success
In October the U.S. Department of Commerce’s Bureau of Industry and Security released an unprecedented set of export controls targeting the semiconductor industry in the People’s Republic of China (PRC). To assess their impact and potential effectiveness, NBR’s Will Nelson interviewed Emily Weinstein, a research fellow at Georgetown University’s Center for Security and Emerging Technology and a nonresident fellow at NBR.
How will the Biden administration’s new semiconductor export controls affect China’s science and technology (S&T) ecosystem? Are there any technologies that will be affected more than others?
In terms of the broad implications, we should not expect to see these controls completely hamper China’s technological development. For example, companies like Baidu have stated that while the controls will have some effect, they are unlikely to be significantly affected, at least right now. The only company that is at major risk of having to shut down is Yangtze Memory Technologies Corporation. Everything else is likely to continue to chug along.
Instead, what we probably see is China facing strategic delays or extra costs in its progress on artificial intelligence (AI). However, this comes with a caveat, given the unilateral, and in some cases extraterritorial, nature of the controls. While the export controls may be effective in the short term, they can become less effective and even counterproductive over time. If the United States wants the controls to remain useful, it needs to multilateralize or plurilateralize them. Otherwise, there are too many ways for China to undermine or circumvent them.
The controls themselves spell out in great detail what kinds of technologies will be affected. For example, the most advanced chips like Nvidia’s A100s are bought specifically because of their processing power for use in big data centers; they are not really commodity chips like those used for dishwashers. So, in that sense—for that tiny piece of technology—the export controls will have a big impact. However, Nvidia has already announced that it will release the A800, which is right below the threshold that would trigger export controls. Essentially, U.S. policymakers need to assume that companies will attempt to find workarounds to preserve access to the PRC’s lucrative market. I am not saying this is necessarily a bad thing. We do want Nvidia to keep making money as long as it does not happen at the expense of U.S. national security.
Overall, I am concerned less about specific technologies than about specific companies. These controls are an important part of competing strategically with China. But the United States also needs to make sure that companies in this country and in allied countries like South Korea, Japan, and the Netherlands are able to remain competitive and push forward in R&D. The United States cannot compete economically with China if these companies are struggling.
How do you think U.S. allies and partners—especially Japan, the Netherlands, and South Korea—will react to these controls? Will the Biden administration be able to gain their support?
Some allies have been more vocal than others. For example, the Netherlands has stated that it will look at the situation according to its own strategic interests and respond accordingly. Overall, none of these countries are optimistic about the Foreign Direct Product Rule (FDPR)—no one wants another country to extraterritorially govern their exports.
However, some countries appear less concerned. Japan has essentially complied and says the change will have minimal impact. South Korea is actually the country that I am most worried about. Combined with the aftermath of the Inflation Reduction Act provisions that exclude South Korean companies from U.S. electric vehicle and battery subsidies, the export controls have not painted the United States in a great light in South Korea.
The United States has given one-year exemptions to the two big South Korean companies—SK Hynix and Samsung—that produce semiconductors in China, mainly for export outside China. The exemptions are largely so that they do not have to completely reshore their production in an impossibly tight time frame. But I would not call it an olive branch either. Rather, it is a sign that the United States recognizes that South Korean firms are in a bind and is trying to show its willingness to work with South Korean partners.
Elizabeth Economy, a senior adviser for China at the Department of Commerce, recently described this dynamic by asking, “Is the U.S. coordinating, or informing its allies?” With allies, the United States essentially took a unilateral approach—similar to how it informs domestic companies like Nvidia of export controls on China—by just handing them a list and saying, “You’re not allowed to export these products.” Economy noted that countries like South Korea have stressed that they wish the export controls had been more collaborative, which is likely causing friction within the alliance.
If the United States wants other countries to implement these policies, it needs to have these difficult conversations before the controls are implemented. Getting allies on board with controls on China will always be difficult—especially for South Korea, given its trading relationship with China. But it is still important to talk, even if coordinating multilateral efforts is not easy. Russia’s invasion of Ukraine was unique in how it galvanized international support for widespread sanctions; such measures are usually not that easy to implement multilaterally.
In this context, it is understandable why the unilateral route was appealing for the Biden administration. It is quicker, and effective in the short run. However, the more the United States continues to rely on unilateral—and especially extraterritorial—controls, the harder it will be to work with allies, or convince allies that the United States will not just turn around and apply extraterritorial controls anyway, even if they do cooperate in certain areas.
Would it be better for the United States to instead hold off on the export controls and keep its leverage over China on strategic or high-tech products? What industrial policy tools does Beijing have at its disposal to respond, and do you think the response will escalate into broader decoupling?
There is an argument among international relations scholars who focus on economic interdependence that maintaining economic interconnectedness reduces the probability of countries going to war with one another, since the costs are too high. However, in the U.S.-China semiconductor context, I think there are ways of implementing export controls that allow the United States to hold onto certain chokepoints in the supply chain. Georgetown University’s Center for Security and Emerging Technology has done some great work on this. Extreme ultraviolet photolithography is the quintessential example of a chokepoint where U.S. allies and partners are strong and China is weak. There is one company in the Netherlands, ASML, that makes these machines, so denying China access to them—which we have already done—is a way to limit China’s ability to develop the most advanced chips.
Pairing that with a decision to grant China access to the more mature machinery that ASML makes (i.e., what is needed to make lower-end, less sophisticated chips) would be a way for a U.S. ally (and by extension, the U.S. alliance network) to retain its strategic chokehold while ensuring that it does not lose too much of its corporate profits. This gets back to my point that key U.S. and allied companies need to remain competitive. The United States cannot just tell ASML not to ship anything to China, otherwise ASML would go out of business. This type of balancing act is important in leveraging technological chokepoints.
However, the other side of the chokepoints is where China is strong and the United States is weak, especially in things like personal protective equipment, pharmaceuticals, and rare minerals. These chokepoints are particularly relevant to understanding China’s response to export controls, given that the country has a much more aggressive and coordinated industrial policy than the United States.
For example, China has even told companies in certain instances to not worry about their bottom line—essentially, “this is in our strategic interest, so we will take care of you.” The fact that China, unlike the United States, is willing to heavily subsidize and dump as much money as it wants into certain industries and countries means that it can squeeze its way into certain supply chains in a way that is kind of parasitic. This is connected to the Chinese concept of “dual circulation,” which holds that through this type of strategic planning China can insulate itself from external supply chain or economic shocks.
Overall, there are pros and cons to remaining economically connected with China. At the highest levels, the conversations around decoupling are pretty solidly against full decoupling, even in strategically vital technologies. For instance, in technologies like AI, the field is highly interdisciplinary. AI technologies can do everything from helping design a hypersonic missile to scanning an X-ray so that a doctor can identify a tumor. If we assume that the United States cannot fully decouple, determining where it is more willing to accept a certain amount of risk to national security provides a way to think of guardrails going forward.
Overall, how would you assess the efficacy of the Biden administration’s policies toward S&T competition with China, and is there anything you would suggest that it do differently?
Foremost, the United States should avoid using the FDPR as much as possible and instead work in consultation with allies. There were probably a number of factors that pushed the Biden administration to opt for the unilateral route and release the export controls when it did, especially the U.S. midterm elections, the G-20 meeting in Bali, and the bilateral meeting between Biden and Xi Jinping.
But again, for the reasons described above, the sooner the United States can multilateralize or plurilateralize those controls with allies and partners, the better off it will be. This is actually written into the new rules, which say that if the United States is able to multilateralize the controls, then it will take away the FDPR and the other unilateral pieces, which is a promising step forward.
In terms of the broader S&T strategy, the Biden administration has done a good job at trying to balance the “promote” and “protect” sides of competition. With chips in particular, the October 7 export control rule is on the “protect” side of competition. But there is also the passage of the CHIPS and Science Act, which emphasizes promoting and incentivizing domestic innovation.
The passage of the CHIPS and Science Act and the export controls close together is a promising first iteration of a quasi-industrial policy in the United States. In particular, the administration’s focus on making the promote and protect sides of economic competition work in tandem could be a winning strategy going forward. I hope this can be a model for other sectors, especially critical ones like batteries, solar power, and energy storage—all technologies that also have complicated or Chinese-sourced supply chains.
However, an important caveat—as I mentioned earlier—is to balance this approach with the considerations of allied countries. In particular, South Korea is concerned that its battery makers have been cut out of the U.S. market by the Inflation Reduction Act, which states that batteries with Chinese components are ineligible for U.S. subsidies. Since China dominates certain parts of the supply chain, all South Korean batteries have some Chinese components in them. This is an example of an instance where batteries have developed as a strategic chokepoint similar to those that have developed in the supply chain for semiconductors. In order to effectively implement legislation like the CHIPS and Science Act and policies like the export controls, the United States needs to take these dynamics into account.
Going forward, even with the challenges for alliance management that these policies pose, it would be interesting to see if the United States could build a strategy, as it did with chips, that both protects and promotes the supply chain for batteries. For instance, along with the export controls targeting Chinese-produced batteries, there could be legislation that offers incentives to facilitate reshoring parts of the global battery supply chain out of China.
This is easier said than done. For example, one significant issue with batteries is that they are heavily reliant on critical minerals like cobalt that must be mined. This has huge social and environmental costs, something that China is much less likely to have qualms about than the United States. Another factor is geography—countries might have access to the critical minerals they need to reshore, or they might not.
Again, the issues raised are complicated, but the United States could use the model of the CHIPS Act plus October 7 export controls to work with its allies on the supply chains for key technologies like batteries and solar panels, all of which are heavily diversified and globalized. Working with allies to come up with a CHIPS 2.0 for other key sectors would be a positive way to move forward and continue to compete with China.
Emily Weinstein is a Research Fellow at Georgetown University’s Center for Security and Emerging Technology and a Nonresident Fellow at NBR.
This interview was conducted by Will Nelson, an intern with the Center for Innovation, Trade, and Strategy at NBR.