What the United States Should Do to Keep Its Lead
he Biden administration’s plan for the United States to control the future of computing was never going to be easy, but at least it was meant to be straightforward: a strategy of denying China access to a limited group of advanced semiconductors while upgrading chip capabilities at home. In August 2022, President Joe Biden signed the CHIPS and Science Act to boost U.S. manufacturing of advanced chips. Two months later, the White House introduced sweeping export controls designed to restrict China’s ability to buy or build the most sophisticated computer processors. Together, these moves aimed to position the United States favorably in a struggle over what is quickly becoming the world’s most consequential industry.
For years, semiconductors have been at the heart of the global economy, and nowhere more so than in China, where chip imports exceed oil imports. Microchips of varying complexity are ubiquitous in everyday devices, but the ultra-advanced chips targeted in Biden’s export controls are unique in their outsize impact. Crucial to expanding cutting-edge supercomputing and artificial intelligence, the chips in question would have played an indispensable role in building out China’s future AI ecosystem, bringing considerable benefits to China’s economy, weapons systems, and surveillance apparatus. By leveraging these chips’ highly concentrated manufacturing pipeline—disproportionately located in countries friendly to the United States—the Biden administration hoped to cut off China’s access to them altogether.
A year after the export controls were announced, Beijing is complicating Washington’s plan. Not only is the Chinese government doubling down on cutting-edge processors, it is also becoming more competitive in legacy chip manufacturing and chip design, threatening to take advantage of U.S. dependencies on Chinese semiconductors. Given China’s response to U.S. restrictions, it is increasingly likely that Washington will have to wage the chip war on two fronts: expanding export restrictions for leading-edge chip production while addressing China’s growing advantage in legacy semiconductor manufacturing and chip design.
TROUBLE IN THE YARD
The Biden administration has said it is adopting a “small yard, high fence” strategy to protect critical technologies. “Chokepoints for foundational technologies have to be inside that yard, and the fence has to be high,” National Security Adviser Jake Sullivan said in October 2022. Understanding what is in this small yard requires translating chipmakers’ jargon regarding “process nodes”—that is, particular chip-building processes that determine how densely designers can pack circuits onto their chips. Roughly speaking, the smaller the process node, the more powerful the chip. The export controls the Biden administration introduced a year ago clamped down on the sales of equipment and materials needed to produce chips with anything smaller than a 16-nanometer process node—ostensibly restricting China from purchasing capabilities to build microchips that were cutting-edge in 2014. The restrictions also prevent China from buying the most advanced chips for supercomputing and AI directly. These measures were so severe that a New York Times article described them as the equivalent of a “declaration of economic war on China.”
Since then, however, China has signaled to the world that it fully intends to survive without access to the American yard. This August, while U.S. Commerce Secretary Gina Raimondo was visiting the country, Chinese technology company Huawei released a phone featuring an unexpectedly advanced processor. The process improvements needed to make this domestically produced chip and the advanced design capabilities on display suggested that China could be gaining ground faster than expected. In terms of chip design, Huawei could soon be on par with global leaders, such as the American technology company Qualcomm. There are caveats to this feat, however: it remains to be seen if China can produce the advanced chip in a cost-competitive manner. The fact that the processor was made on a 7-nanometer process node, which relies on a capability that U.S. export controls seek to deny, was not surprising given that China’s top chipmaker, SMIC, has been producing similar devices since 2022. Furthermore, China is unlikely to perfect extreme ultraviolet (EUV) lithography, the advanced technology needed to wield process nodes under 5 nanometers (the state of the art in 2020), for some time.
Despite these caveats, one thing is certain: Chinese companies will continue trying to vault over and tunnel under Biden’s high fence by any means necessary to replicate the capabilities being kept inside. That means accelerating the creation of indigenous supply chains with state-backed capital, finding alternative routes to similar capabilities, or illicitly procuring restricted technology. Already, Huawei has been found to be indirectly siphoning Taiwanese firms’ crucial chip-making expertise through a discreet network of affiliated chipmakers. China’s next-generation lithography tool—which uses a form of lithography one generation shy of Dutch industry leader ASML’s most advanced technology—is set to be launched this year, demonstrating China’s diminishing dependence on foreign chip-making tools. And Chinese companies such as YMTC, the country’s largest memory chipmaker, are making progress in producing advanced semiconductors using domestic equipment. The upshot is likely going to be an expanding list of parts and tools that require further restrictions from the United States and its partners to keep China from building out its domestic capabilities. Walling off even more technologies will mean difficult tradeoffs between economic and security interests.
Chinese ambitions have so far outpaced American concerns.
China has already demonstrated that it intends to become a global leader in advanced semiconductor technology. In 2021 and 2022, Beijing undertook an anticorruption drive in its semiconductor funds, punishing individuals for graft and wasteful spending, a move that managers on the ground believe was driven by a frustration over their failure to deliver on strategic objectives. But instead of shutting down the funds, the Chinese Communist Party reinstated some business leaders and expanded its Integrated Circuit Industry “Big Fund”—the large, centralized pool of cash intended to help China achieve semiconductor self-sufficiency—with a clearer focus on overcoming Western manufacturing chokeholds. The first iteration raised $19 billion, the second raised $27 billion, and the up-and-coming third round of fundraising is set to be the largest yet at $40 billion, although China’s flagging economy may throw a wrench in that ambitious goal. The country also launched a new Central Science and Technology Commission in March, a body that helps provide major companies such as SMIC, Hua Hong, and Naura direct access to funding and state-backed research.
These initiatives suggest that although its state-led sprints can be grossly inefficient, Beijing is learning from experience, adapting its policies, and exhibiting a continued willingness to double down on chip-making. This does not mean the inefficiencies will disappear entirely, but as private-sector and party leaders correctly insist, their efforts need not be perfectly efficient to be ultimately effective.
A BATTLE ON TWO FRONTS
Addressing Beijing’s attempts to catch up on leading-edge technology is difficult enough on its own, but now, the United States is seeing challenges crop up in new areas of chip-making. Industry and government leaders in China see the struggle with the United States as a two-front battle, not just for the cutting-edge semiconductors used in phones and AI accelerators but also for what are known as “legacy” chips, which are manufactured using older technology. And in contrast to the state-led campaign to develop super-advanced chips, China’s push for legacy chip supremacy has the advantage of being propelled by an even more potent force: the gargantuan Chinese consumer market.
The term “legacy” refers to semiconductors produced on older process nodes with less dense circuitry, but the label can be misleading. Legacy chips are used in almost everything other than the most advanced computer processors. They form the backbone of critical infrastructure, weapons systems, and industrial equipment. This category also includes state-of-the-art microwave integrated circuits and compound semiconductors, which enable technologies from 5G communications to electric vehicles.
A future in which China dominates the legacy chip ecosystem would give the Chinese government enormous leverage. Beyond economic advantages (recall how important legacy chips were for automakers during the COVID-19 pandemic), it could allow China to turn off the servicing or supply of parts critical to essential systems such as U.S. energy infrastructure and open backdoors for conducting espionage. This reality is already starting to dawn on American and British policymakers, who are realizing that China could use chips inside smart devices to spy on or even control smart cars, emergency response systems, and everyday consumer electronics. Last year, Chinese companies captured more than three-quarters of the global market for the subset of legacy chips that power smart device connectivity.
Although some U.S. analysts have considered these threats, Chinese ambitions have so far outpaced American concerns. Beijing’s subsidies have led to a large supply of chip capacity: SMIC, one of the biggest recipients of Chinese state funding, has been ramping up legacy manufacturing and is currently booked solid with orders. Its new factories also have unique advantages. They produce legacy node chips in batches on large, 12-inch-diameter silicon wafers, enabling them to print more chips at a lower cost than the 8-inch equivalents prevalent in the United States, and use ever-larger proportions of Chinese-made equipment that is cheaper to acquire.
The United States is seeing challenges crop up in new areas of chip-making.
China’s expansion in semiconductor manufacturing is poised to hurt one other aspect of the American semiconductor ecosystem: chip design. Design has long been an area of U.S. strength, with the United States’ lead so pronounced that little attention was paid to design issues in the CHIPS Act. Nonetheless, here, too, China has a growing comparative advantage. Since China is the single biggest consumer of semiconductor devices in the world, its design firms are well-placed to build customer relationships that make use of the country’s large captive market. Uncertainty about Western sanctions has led many Chinese buyers, previously unwilling to order from scrappy startups, to turn to domestic chipmakers. With lower labor costs and a rapidly expanding talent pipeline, China is well on its way to designing increasingly more advanced chips at home.
Already, swarms of Chinese startups are taking advantage of China’s cheap engineering talent and its growing legacy chip manufacturing capacity to design a vast array of processors, modules, and sensors. In the process, they are slowly creeping up to higher-end products that can compete with those produced by American firms. Here again, Huawei’s August phone launch does not bode well for the United States. It demonstrated that the company has successfully mimicked the ability of leading U.S. tech firms to design remarkably sophisticated chips, albeit with lingering dependencies on Western software and intellectual property.
This more market-driven path of industrial development is distinct from the state-led leapfrogging that China is trying to accomplish when it comes to leading-edge chips. It more closely resembles China’s successful ascent in telecommunications equipment, solar panels, and electric vehicles. As in those sectors, China’s market-led boost in manufacturing could help it master chip design and challenge U.S. dominance. Given Chinese business leaders’ enthusiasm for legacy chips and the market forces in their favor, there is little reason to doubt that China will make dramatic progress in expanding its market share in both legacy chip manufacturing and design.
What Washington should do about these developments is the subject of intense debate. The overall effects of the CHIPS Act and the Biden administration’s export controls are still uncertain, leading some observers to question the efficacy of a commercial sanctions strategy altogether. But there is also the opposite temptation: to expand the fence beyond leading-edge chips to try to address China’s growing dominance in manufacturing legacy chips. But the vast complexities of the legacy market and its lack of clear chokepoints make a high fence approach infeasible. Such a plan would escalate distrust in international markets, lead the United States to lose more leverage over China’s tech sector over time, and amplify incentives for Chinese players to charge ahead.
The best policy is to instead recognize that leading-edge manufacturing, legacy manufacturing, and chip design each require unique responses. Whereas the United States has good reason to deny China access to specific cutting-edge processors, the United States need only ensure that it does not become dependent on Chinese products and stays competitive in chip design.
For the most advanced technologies, the small yard, high fence approach still has a lot of mileage, as long as the fence is continually maintained. Indeed, in light of Huawei’s progress, export bans for semiconductor manufacturing equipment are more relevant than ever. The U.S. Department of Commerce should continue to closely monitor China’s leading-edge chip manufacturing efforts for sanctions loopholes. It should also zero in on China’s efforts to develop advanced lithography methods, including EUV and ion beam lithography, by securing the United States’ and allies’ supply of upstream components used in the development and production of these technologies. Given the speed of adaptation of Chinese companies, it is better to err on the side of excess restrictions in targeted areas than risk having insufficient measures. In addition to choking off China’s ability to produce advanced chips, the United States should also reduce the demand for China to make them in the first place. This can be done by developing ways to limit the uses of advanced chips for dangerous AI applications, such as built-in hardware mechanisms that prevent chips from being used for prohibited purposes. This would make it possible for the United States to sell leading-edge chips to China for legitimate uses while preserving its leverage over the production and supply of advanced semiconductors.
The small yard, high fence approach is not defunct.
For the manufacturing of legacy chips, a different approach is needed, as the United States must only ensure that it has suppliers other than China in critical areas. And unlike China, the United States does not have to go it alone. It should work to diversify supply chains with friendlier countries such as India and Mexico, both of which have a growing comparative advantage in production and packaging. The U.S. government should also empower the Federal Communications Commission to more thoroughly scrutinize foreign components for backdoors that could be used for spying and set up a working group at the U.S. Department of Homeland Security to evaluate critical dependencies on legacy chips.
Finally, in chip design—an area left largely unaddressed by the CHIPS Act—the U.S. government should work to amplify the country’s comparative advantages. The Electronics Resurgence Initiative, a program launched by the Pentagon’s Defense Advanced Research Projects Agency in 2018 to reinvent semiconductor design using automated tools, offers an example of how the government can invest in ambitious innovation that puts the U.S. private sector ahead. Furthermore, American companies could work together to bolster the software and standards built on American chip designs, investing in interoperability and shared platforms to strengthen the network effects that make the U.S. chip ecosystem best of class. Doing so could stave off Chinese advances and ensure that the United States retains its lead in designing the digital infrastructure that will undergird tomorrow’s economy.
Contrary to the hype and concern about Huawei’s new capabilities, the small yard, high fence approach to Chinese semiconductors is not defunct. But to respond to China’s most recent moves, the United States must patch the holes in its fence while crafting new methods to outpace Beijing—and fast.