A major, clandestine production run by TSMC for Huawei shell companies supplied China with millions of advanced AI chips. This single violation artificially propped up China's AI compute capacity, effectively delaying the full impact of U.S. export controls by two years and obscuring the true state of China's domestic capabilities.
Related Insights
Reports of China building a working EUV lithography machine are misleading. The effort appears to be an assembly of smuggled components from ASML's existing supply chain, not a story of domestic innovation. This frames the primary challenge as one of export control evasion rather than a rapid technological leap by China.
The decision to allow NVIDIA to sell powerful AI chips to China has a counterintuitive goal. The administration believes that by supplying China, it can "take the air out" of the country's own efforts to build a self-sufficient AI chip ecosystem, thereby hindering domestic firms like Huawei.
The most significant sanctions loophole isn't physical chip smuggling but 'compute smuggling.' Chinese firms establish shell companies to build and operate data centers in neutral countries like Malaysia. They then access this cutting-edge compute power remotely, completely bypassing physical import restrictions on advanced hardware.
The "Operation Gatekeeper" bust uncovered a massive illegal AI chip smuggling operation into China. This indicates that prior to the recent policy change, a significant black market existed to circumvent US export controls, suggesting high, unmet demand that official numbers don't capture.
While NVIDIA laments lost revenue from export controls, those same policies blocked its primary Chinese competitor, Huawei, from accessing TSMC's advanced manufacturing. This prevented Huawei from launching a competing 7nm GPU, preserving NVIDIA's market dominance in China.
Contrary to their intent, U.S. export controls on AI chips have backfired. Instead of crippling China's AI development, the restrictions provided the necessary incentive for China to aggressively invest in and accelerate its own semiconductor industry, potentially eroding the U.S.'s long-term competitive advantage.
China's superior ability to rapidly build energy infrastructure and data centers means it could have outpaced US firms in building massive AI training facilities. Export controls are the primary reason Chinese hyperscalers haven't matched the massive capital spending of their US counterparts.
The U.S. government identified a critical loophole allowing Huawei to acquire advanced High Bandwidth Memory (HBM) but waited nearly a year to close it. This bureaucratic delay, from February to December 2024, provided a significant window for China to stockpile essential components, undermining the broader export control strategy.
China is engaged in a strategic propaganda campaign, exaggerating its technological self-sufficiency in areas like AI chips. The goal is to convince U.S. policymakers that export controls are futile. This narrative aims to pressure the U.S. into relaxing restrictions, which would then allow China to acquire the very technology it claims not to need.
U.S. export controls on advanced semiconductors, intended to slow China, have instead galvanized its domestic industry. The restrictions accelerated China's existing push for self-sufficiency, forcing local companies to innovate with less advanced chips and develop their own GPU and manufacturing capabilities, diminishing the policy's long-term effectiveness.
