New Export Controls on Chinese Semiconductors May Prove Self-Defeating

In July, news broke that China had made a two-generation leap in semiconductor development, despite U.S. export controls on advanced chipmaking technology. According to Tech Insights, Semiconductor Manufacturing International Corporation (SMIC), the leading Chinese chipmaker, was able to produce 7-nanometer (nm) semiconductors using tools designed to fabricate chips at the 14nm node and above.

This development has unsettled observers of China’s semiconductor industry who worry that SMIC’s manufacturing capabilities now outstrip those of most American producers and are approaching those of leading global foundries. Although node size is only one variable among many, China’s new 7nm chips could enable the production of more advanced electronics, including everything from supercomputers to state-of-the-art military equipment. In the weeks following the release of Tech Insights’ report, the Department of Commerce notified all U.S. semiconductor equipment manufacturers that it would widen export bans on gear used to make anything smaller than 14nm chips, including tools that enabled SMIC’s breakthrough; it appears that Washington is shifting from a strategy of slowing down China’s future tech development to degrading its existing capabilities.

But these new restrictions are not only likely unnecessary, they may ultimately prove self-defeating. Overly zealous controls that limit older semiconductor equipment sales to China will inflict collateral damage on American, and potentially international, semiconductor firms, and may not significantly damage Chinese firms’ ability to innovate. So rather than enacting blanket bans on existing chipmaking technologies, the United States should focus on pushing the technological frontier forward.

The calls for more stringent export controls fail to address a core reality: that SMIC’s achievement may not be quite as momentous as it seems. Some analysts have hailed the 7nm chip as a breakthrough for China’s semiconductor industry. Instead, it points toward Beijing’s continued inability to achieve self-sufficiency in a highly specialized and globalized industry. Questions remain about the new chips’ yield, memory capacity, and energy efficiency, as well as whether the company can manufacture them at scale. This is largely because SMIC was forced to use suboptimal manufacturing techniques with decades-old deep ultraviolet (DUV) manufacturing equipment. In short, not all 7nm chips are created equal.

Why must SMIC rely on outdated DUV technology? Because the United States already has in place export controls on the most advanced chipmaking tools. That is: existing U.S. export controls appear to be working. SMIC’s latest “breakthrough” may very well be its last.

More broadly, while Chinese companies increase their chip production, they remain unable to compete with the Taiwanese, South Korean, and American companies that produce most of the world’s cutting-edge semiconductors. This is because, due to effective export controls, China lacks the extreme ultraviolet (EUV) lithography software and equipment that have fueled the latest leaps in chip technology. Beijing has some advanced technical know-how, but it does not have enough of the best design and manufacturing tools to become globally competitive. Increasingly frustrated with the lack of progress, the Chinese government has launched investigations of several officials at the “Big Fund,” a state-led venture capital fund tasked with driving semiconductor innovation.

Despite the limitations of Beijing’s chip industry, Washington has responded to China’s chipmaking ambitions with increasing alarm. In 2018, Congress passed the Export Control Reform Act (ECRA), mandating that the Department of Commerce identify and control “emerging and other types of critical technologies” that “are essential to the national security of the United States.” Since then, export controls on chips and chipmaking technology have broadened. Between 2018 and 2022, the number of Chinese companies on the Commerce Department’s entity list—a group of end-users to whom exporters must first obtain a license to sell controlled technology—quadrupled to nearly 600. The number of export licenses rejected by the Bureau of Industry and Security also rose from 177 to 526 between 2020 and 2021.

Recently, the U.S. Department of Commerce has gone one step further by notifying U.S. semiconductor equipment makers that it would restrict the sale of tools used to manufacture sub-14nm chips. Under one proposal, Commerce would limit exports of chipmaking tools to Chinese foundries that make chips under 14nm, while allowing exports to less advanced production facilities, even those operated by the same parent company. Some Republican Congressmembers have argued that these restrictions should apply to exports to all foundries in China. In practice, Commerce has already declined licenses at the 14nm level on a case-by-case basis, but the institutionalization of sub-14nm controls would represent a major escalation in the U.S. export control regime.

Yet, given Beijing’s continued inability to compete with the most cutting-edge semiconductor manufacturers, dialing up the bans in the wake of SMIC’s 7nm “breakthrough” is not sound policy. Selective protection of U.S. technology is necessary for the most advanced chipmaking technology, but broad bans on older technology may do more to hurt U.S. semiconductor firms than to protect U.S. national security.

This is because China is the largest market for semiconductor equipment, providing revenue streams for American companies to pile their profits into the R&D that fuels innovation. Estimates suggest that the prospective ban on sub-14nm manufacturing equipment to China could impact 25-30 percent of Applied Materials’ sales, a top U.S.-based manufacturer of chipmaking equipment. Historically, chipmakers have cut back on R&D spending during periods of declining revenue.

Second, due to the globalized nature of the semiconductor industry, the United States will need to bring allies and partners on board before implementing more expansive controls. Otherwise, U.S. companies will be unable to sell to their largest market, while China continues to buy the tools it needs from other foreign suppliers. Coaxing South Korea, Japan, and the Netherlands into accepting new U.S. controls may prove difficult, as their semiconductor equipment manufacturers would lose billions in sales. If pushed too far, leading chipmakers could decide to establish new foundries without American technology to skirt expansive U.S. controls.

Whatever the configuration of export controls, however, Washington should expect that China will maintain its long-standing emphasis on indigenous innovation. During a famous speech to the Chinese State Council in 1958, Mao Zedong declared that the total trade embargo imposed by the United States during the Korean War had forced Beijing “to work out all the solutions” on its own. Over the course of the country’s modern development, China has been surprisingly good at achieving technological advances without foreign help.

Both official policy and geopolitical realities have shaped China’s drive toward technological “self-reliance.” Despite being cut off from Soviet missile and nuclear technology in the 1960s, Beijing succeeded in developing both sectors indigenously, and now produces some of the world’s most advanced missiles. The country’s space capabilities also improved after the United States ceased cooperating with the Chinese space industry in the 1990s. More recently, Beijing has marshaled its impressive industrial resources to make advances in AI, 5G, and energy storage, among other technologies.

For this reason, one theory holds that U.S. export controls could spur a breakthrough in China’s semiconductor industry. But thus far, Beijing’s experience with semiconductor development has followed a different pattern. In 2015, the State Council set out to supply at least 70 percent of China’s demand for chips with domestic production by 2025. China has tried to wean itself from reliance on foreign products by pouring subsidies into chip research, design, and production, pilfering industrial secrets, and recruiting elite international talent. Yet despite investing billions in the semiconductor industry, China controls less than one percent of electronic design software and semiconductor tools and materials.

All of this means that the U.S. should stop wasting its time on restricting the export of older, less critical technologies. Instead, Washington should focus on pushing the technological frontier forward. The recently passed CHIPS and Science Act is a step in the right direction, but its investment in domestic human capital is insufficient; the bill only set aside U.S.$200 million for semiconductor workforce training in the United States. China’s ability to attract foreign talent, on the other hand, including top engineers from foreign tech firms, has been an ignored yet crucial component of the limited innovations it has made. The United States, for its part, has benefited greatly from attracting leading global science and technology talent. Export controls alone do little to push forward human capital development in the United States, nor do they hamstring China’s access to international talent.

A better approach would be to maintain the “small yard, high fence” approach that commentators and experts have been touting for several years. Washington should focus on protecting only the most advanced dual-use technologies with airtight, multilateral controls, while allowing other forms of trade and commerce that benefit the U.S. tech ecosystem. As it stands, the Biden administration is set to carry forward U.S.-China decoupling on a larger scale than did Trump, with potentially deleterious effects on the innovative capacity of domestic chip producers.