The development and production of leading-edge semiconductors is a difficult, expensive, and risky endeavor. Case in point: Intel, once the world’s top chipmaker, has fallen behind Taiwan-based TSMC and Korea-based Samsung in mass-producing the latest, most powerful generation of chips, used in applications such as artificial intelligence, autonomous vehicles, and high-end weapons systems.
Moreover, each new generation of chips requires comparable advances in related technologies like software for designing chips (primarily made by U.S.-based companies such as Cadence or Synopsys) and ultra-precise photolithography equipment from companies like Netherlands-based ASML and Japan-based Nikon that are essential to the chip-making process.
So far, there are no Chinese companies on the very short list of global businesses contributing key technologies to leading-edge advanced chips, which aren’t easy even for companies like Intel. However, Chinese manufacturing facilities have become key producers of “mainstream” or “legacy” chips, which are powerful but “not-the-latest-model” chips that run everything from appliances to tire sensors on cars.
Against the background of this complex global chip ecosystem, the United States is pursuing a clear policy: To “maintain as large of a lead as possible” over competitors in “key technologies,” in the words of President Biden’s National Security Advisor, Jake Sullivan. In particular, this goal reflects mounting concern in Washington about China’s long-term strategic threat, both military and economic. Notably, the ruling Chinese Communist Party is doubling down on a “military-civil fusion” research and development strategy that links civilian and military modernization and technology development. The immediacy of the threat can be argued, but not its nature.
The new policy has two parts and a difficult choice. The first part is to slow down the transfer of Western knowledge and capabilities to the Chinese semiconductor industry. This goes under the name “export controls,” but it really means a series of actions, described below, aiming at slowing down or stopping the conveyor belt that for the past two decades has swiftly and efficiently carried technology from the Western countries to the broad swathe of thriving Chinese factories, including those supplying the Chinese military.
The second part of the new policy must be aimed at maintaining and even accelerating advanced Western chip progress. The CHIPS Act is a start, but as noted earlier, the cutting-edge of semiconductor technology is a dangerous and expensive place to be. The U.S. is still the leader, but it can’t expect to go it alone. Other countries have unique knowledge and capabilities that must be appreciated and respected. For example, ASML is the only company in the world making extreme ultraviolet (EUV) lithography equipment, without which the most powerful chips would not be possible.
The difficult choice is how far to extend the new controls in practice. There is a general consensus (though in some cases grudging) that China should be denied the latest and greatest chip technologies. To that end, the Commerce Department’s Bureau of Industry and Security (BIS) is prohibiting the transfer of advanced semiconductor technologies to any Chinese company or fab that manufactures chips for supercomputers and other military applications is included. Moreover, American workers and companies that operate or service advanced chipmaking equipment in China henceforth will require a special waiver to continue their work. The new controls also require non-U.S. companies to comply with the new restrictions or risk being cut off from using American semiconductor technology. Indeed, the U.S. was able to convince the governments of the Netherlands and Japan, home to key suppliers, to ban the export of leading-edge technologies.
This policy, directed toward advanced chips, could well be successful in slowing China’s semiconductor progress. No matter how much China spends on semiconductor R&D, it may not be able to easily match the combined intellectual and scientific heft of the U.S., Europe, Japan, Taiwan, and Korea working together.
But the consensus begins to fray once the policy conversation turns to putting similar controls on mainstream or legacy chips. On the one hand, mainstream semiconductors are still very powerful in historical terms, and the technologies used to make them can be potentially adapted to produce more advanced chips. An argument can be made that to be truly effective, controls have to be both broad and tight. One example of potentially tight controls: The Commerce Department just proposed a set of “guardrails” that would prohibit recipients of CHIPS Act funds from expanding the capacity of their existing legacy manufacturing plants in China by more than 10% from today’s level. That would effectively put a permanent cap on U.S. production of legacy chips in China for export-oriented products.
On the other hand, these chips, used by the bucketful in all sorts of modern applications, are low-margin, and therefore a low-cost country like China logically has a comparative advantage. Moreover, automakers, aerospace companies, and other U.S. businesses benefit from the ability to import these chips at a low price, which then in turn benefits U.S. consumers.
How to escape this conundrum? The key is to look ahead toward the future. The technical bar for each new chip generation gets higher and higher, so the more hands the better when it comes to pushing the frontier forward. The Semiconductor Industry Association estimates that meeting future semiconductor needs will require $3 trillion in investments over the next decade for R&D, supply chain resiliency, and new fabs, both for advanced and mainstream chips. To put this in perspective, Intel’s total capital investment and R&D budget in 2022 came to $42 billion, suggesting what is needed is about 10 Intels.
On a global basis, that’s doable. What’s needed is a “coalition of growth” with no single country monopolizing the funding, expertise, blueprint, or supply chain for manufacturing semiconductors. This diffusion of technical know-how necessitates a balanced approach to address national security needs, safeguard semiconductor supply, and advance semiconductor technology and equipment, as the Department of Commerce and BIS are doing by soliciting public comment on the new restrictions.
The importance of semiconductors will only continue to rise in a digitized society. No one disputes the importance of semiconductors for national security. The United States cannot compromise on ensuring a secure global environment. However, the essential nature of chips will require a roadmap that integrates national security needs with chip advancement and innovation.