Chinese military bodies, state-run AI research institutes, and universities have discreetly procured small batches of Nvidia semiconductors. This happens despite the U.S. ban on exports of advanced computer chips to China.
The transactions, orchestrated by relatively obscure Chinese suppliers, shed light on the ongoing challenges faced by Washington in impeding China's access to cutting-edge U.S. chips, crucial for advancements in AI and sophisticated military computing.
Chinese Access to Advanced U.S. Chips Continues
Contrary to assumptions, the purchase or sale of high-end U.S. chips is not illegal in China. Publicly available tender documents disclose numerous instances where various Chinese entities acquired Nvidia semiconductors, including the A100 and the more potent H100 chip. These chips faced export bans to China and Hong Kong in September 2022. The less powerful A800 and H800 chips, developed by Nvidia specifically for the Chinese market but banned in October, were also among the purchases.
According to Reuters, Nvidia's graphic processing units (GPUs) are widely acknowledged for their superior performance in AI work, efficiently processing substantial data volumes essential for machine-learning tasks.
Despite attempts to foster local alternatives by companies like Huawei, the demand for and access to banned Nvidia chips persist, underscoring the absence of viable substitutes for Chinese firms. Before the bans, Nvidia held a staggering 90% share of China's AI chip market.
Nvidia's GPUs Remain Coveted in China
The purchasers include prestigious universities and entities subject to U.S. export restrictions, such as the Harbin Institute of Technology and the University of Electronic Science and Technology of China.
Both entities have faced allegations of involvement in military affairs, contradicting U.S. national interests. The resilience of this supply chain raises questions about how the Chinese suppliers managed to obtain Nvidia chips, as the company and its approved retailers were not involved.
Despite the creation of an underground market for these chips in China, fueled by excess stock or imports through local entities in countries like India, Taiwan, and Singapore, U.S. export restrictions face inherent challenges.
According to Biz Crast, a professor at Tufts University, the nature of chips allows for easy smuggling, making it unrealistic to achieve watertight control.
The primary objective appears to disrupt China's AI development by hindering the construction of large clusters of advanced chips capable of training AI systems. In this complex landscape, Chongqing University's recent tender for a brand-new A100 chip further emphasizes the tenacity of China's pursuit of these critical semiconductor components.
Photo: Specna Arms/Unsplash