U.S. chip export controls face loopholes, may be propelling China’s self-reliance drive

Two years after Washington moved to wall off China from cutting-edge semiconductors, the strategy faces mounting loopholes and workarounds that have blunted its impact—and may be accelerating Beijing’s push for technological self-reliance.

The Biden administration’s 2022 export controls grew out of national security concerns tied to Chinese companies such as Huawei, vulnerabilities exposed by pandemic-era chip shortages, and efforts to preserve U.S. leadership in semiconductor technology. Huawei’s expanding role in 5G networks and U.S.

reliance on Chinese technology stirred worries about influence over critical infrastructure, while shortages underscored how dependent the U.S. economy had become on Taiwan Semiconductor Manufacturing Company (TSMC), located just across the Taiwan Strait from China.

Both Washington and Beijing view advanced chips as central to future power, underpinning artificial intelligence, advanced weapons, and high-performance computing such as quantum. On October 7, 2022, the U.S. Bureau of Industry and Security announced the most sweeping restrictions in decades.

The rules targeted four areas: advanced AI processors, semiconductor design, fabrication capabilities, and access to manufacturing equipment. NVIDIA’s A100 and H100 graphics processing units were barred from export to China, and U.S. equipment makers including Applied Materials, Lam Research, and KLA Corporation could no longer ship their most sophisticated tools to Chinese firms.

The immediate aftermath was disruptive. Projects around the world were delayed, costs rose, and companies scrambled to adapt. As the measures took effect, however, technical, legal, and enforcement gaps emerged. Regulators’ initial focus on chip specifications—such as interconnect speeds and performance thresholds—left room for firms to design around limits while still delivering near-advanced capabilities.

Some foreign companies stockpiled chips before the rules took effect, creating buffers that eased the transition, and others structured deals narrowly within compliance, exploiting ambiguities in the language. Unauthorized channels also undercut the effort.

Smuggling networks and third-party intermediaries facilitated the flow of restricted GPUs into China, and reports documented that NVIDIA’s A100 and H100 chips were still available for purchase on Chinese e-commerce platforms months after the ban.

The experience echoed past technology-denial attempts—from Cold War-era supercomputer restrictions to recent sanctions on Russia—suggesting export controls can slow but rarely stop determined rivals. China’s scale, resources, and state coordination have made it adept at adapting to and circumventing U.S.

restrictions. Rather than crippling China’s innovation capacity, the controls appeared to galvanize a national drive for technological independence. For U.S. policymakers, the challenge now is closing gaps and enforcing the rules without triggering unintended consequences that erode their long-term effectiveness.