The Gate

From smartphones and data centers to electric vehicles and industrial automation, semiconductors are the foundational technology that enables almost every sector of the modern economy. With fear of China’s economic and military might on the rise, the Commerce Department’s Bureau of Industry and Security (BIS) has taken extreme measures to restrict China’s domestic semiconductor industry over the last several years. BIS first issued export controls on advanced computing and semiconductor manufacturing items on October 7, 2022. A year later, they followed up with a set of even stricter controls targeting a wider set of chips, cracking down on the use of subsidiaries to circumvent the restrictions and blacklisting new Chinese firms from importing advanced U.S. technology. In December 2024, BIS expanded these restrictions even further by implementing controls on 24 new types of semiconductor manufacturing equipment and adding restrictions on high-bandwidth memory chips crucial for AI applications. BIS also introduced controls on three additional types of semiconductor development software and added 140 Chinese companies to its Entity List, which significantly broadened the scope of its export control regime.

Semiconductors are considered a “dual-use” technology since they have both civilian and military applications. The potential military applications of semiconductors are often used as the primary rationale for U.S. export controls. Secretary Gina Raimondo cited them in her explanation of the new controls in 2023, saying that the U.S. should inhibit China’s access to “advanced semiconductors that could fuel breakthroughs in artificial intelligence and sophisticated computers that are critical to [Chinese] military applications.” 

For most goods and technologies, understanding their military application is relatively straightforward. They are directly used to manufacture a weapon or are used in the weapon, and therefore the U.S. should restrict their sale to adversaries. For semiconductors, the story is slightly more complicated. While chips are used in nearly every piece of weaponry and technology in a modern military, the U.S. has only restricted exports on the highest performance chips and the equipment used to make them. Traditionally, export controls on the cutting edge of semiconductor technology have been invoked based on fear that supercomputers using advanced chips can be used to simulate nuclear reactions or train AI models with military applications. 

The U.S. President is currently authorized to issue export controls for dual-use goods and technologies under the Export Controls Reform Act (ECRA) of 2018, but the history of U.S. export controls on dual-use technologies goes back as far as the Second World War. In 1940, Congress authorized the President to control the export of military equipment and munitions, and then, following the Japanese attack on Pearl Harbor, they approved the same controls for civilian goods. While Congress began to roll back strict wartime export controls after 1945, the change was short lived, as the start of the Cold War led to a renewed focus on export control policy. In the late 1940s, dual-use goods and technologies became the primary focus of export controls through a series of initially temporary, but continuously renewed bills like the Export Control Act of 1949. Although the fear of dual-use technologies landing in the hands of adversaries was the primary motivation for export controls, Congress also cited the protection of the domestic economy as a key reason for the new legislation. As a result, U.S. export controls have always involved a complicated interplay of concerns for both the economy and U.S. national security.

While the U.S. leads in a few key areas early in the semiconductor supply chain like electronic design automation (EDA) and IP, it does not have a large enough market share on its own to create a major choke point further downstream in the fabrication, assembly and testing stages. Policymakers and diplomats in the White House and Commerce Department have increasingly relied on key allies including Japan, the Netherlands, and Germany, to close the many gaps in the U.S. export control strategy. In June 2024, U.S. export policy chief Alan Estevez conducted high-level visits to both The Hague and Tokyo, pressing for tighter restrictions on semiconductor manufacturing equipment exports to China. While these countries are strong U.S. allies and are generally wary of the rise of China’s homegrown semiconductor industry, their domestic semiconductor firms rely heavily on sales to China, meaning export controls can be enormously unpopular.

ASML, the Dutch photolithography giant, is the world’s sole supplier of the extreme ultraviolet (EUV) lithography machines required to fabricate the most advanced chips. In Q2 of 2024, almost 50% of ASML’s sales came from China, demonstrating both the company’s reliance on Chinese customers and the Chinese semiconductor industry’s frantic stockpiling of equipment as even stricter controls loom on the horizon.

However, experts are not unanimous in their assessment of the impact of export controls on U.S. and allied semiconductor firms, many of which, like ASML, rely heavily on sales to China. A recent report from CSIS argues that many U.S. firms have lost and failed to replace Chinese customers in the wake of the first round of export controls in 2022. The authors claim that the restrictions have created a hostile environment which has interfered with the formation of productive business relationships between firms. However, during the same period data from U.S. firms producing wafer fabrication equipment shows that gross margins have grown, R&D spending remains relatively stable and sales to China of unrestricted products (and restricted products through unresolved loopholes) remains strong.

In the face of enormous U.S. pressure, China has turned inwards to develop self reliance in all stages of the semiconductor supply chain. In May 2024, the third phase of the “Big Fund,” formally known as the “China Integrated Circuit Industry Investment Fund,” was launched with a pool of $47.5 billion for the purpose of financing to domestic semiconductor champions. Given how much attention and funding China has directed towards the development of its domestic semiconductor industry over the last few years, it is fair to ask whether U.S. export controls only hastened an inevitable decoupling. 

Questions can and have been asked of their effectiveness as well. Despite BIS’s recent updates patching the most egregious loopholes, Chinese firms have proven themselves incredibly adept at evading export controls. In some instances, Chinese fabs producing cutting edge chips which are normally subject to U.S. restrictions have been able to sidestep regulators by using the advanced machinery in nearby legacy fabs which are not subject to the same restrictions. 

SemiAnalysis, a prominent blog covering the semiconductor industry, identified one such case in which the two fabs were connected by a “wafer bridge,” allowing them to effectively function as one giant fab. However, technically only “one building is entity-listed by the U.S. and working on advanced logic for AI chips, a clear national security concern. The other is free to import ‘dual use’ tools as it runs only ‘legacy processes.’” These examples show how evading U.S. sanctions and export controls is not always the complex task it might seem to be. As the SemiAnalysis authors ask, “do you believe they aren’t sharing anything over the wafer bridge?”

As a result of the proliferation of export-control compliant variants of restricted chips and a complex network of distributors working to meet the massive domestic demand, China has managed to keep GPU costs affordable despite U.S. export controls. In September of 2024, the Financial Times reported that the cost of renting an Nvidia chip was cheaper in China than in the U.S. However, it is important to note that the export-control compliant versions of Nvidia’s chips will only lag further behind the cutting edge in years to come, so the impact of U.S. export controls in China may grow stronger as time goes on.

As the U.S. enters the first year of the second Trump administration in 2025, semiconductor export controls will likely become even more stringent. The incoming administration has signaled its intention to take an even harder line on technology transfers to China, with proposed measures including a large blanket tariff on Chinese goods and expanded restrictions on semiconductor equipment and chip exports. These developments suggest that the technological decoupling between the United States and China, particularly in the semiconductor sector, will likely only accelerate over the next four years.

The image used in this article is licensed for reuse under the CC0 1.0 Universal Deed.