Key Highlights:
- Chinese AI firms are seeking to secure Nvidia-powered computing capacity at overseas data centres amid concerns that Washington could tighten restrictions on remote access to advanced chips.
- A proposed US bill, the Remote Access Security Act (RASA), could give the Commerce Department authority to regulate cloud-based access to restricted technology.
- Fears of stricter enforcement are also driving stockpiling of restricted Nvidia AI chips in China’s grey market.
Chinese AI firms plan to gain access to Nvidia-powered computing capacity from overseas data centres. This is following potential US restrictions on remote access to advanced processors. The companies fear that a controversy involving Nvidia CEO Jensen Huang could add to the rift and widen the gap in current export controls. While the US limits export of the advanced chips, Chinese AI firms can acquire them by renting capacity from offshore data centres. The developments have also increased concerns about stringent enforcement against grey-market chip supplies.
Why Are Chinese AI Firms Seeking Overseas Nvidia Computing Capacity?
U.S. export controls hinder the sale of Nvidia’s most advanced AI processors to China. However, Chinese companies can acquire these chips by renting computing capacities at offshore data centers. This arrangement has helped AI developers to use advanced processors without importing the physical hardware into China. As Reuters reported, a major Chinese cloud provider has discussed acquiring overseas Nvidia chip capacity, while some companies have approached data center operators to rent computing resources.
Smaller AI labs are also planning similar arrangements, including dealing with major providers of rented computing power in Southeast Asia. The companies believe that Washington could limit this access route, making offshore computing more harder to acquire. The concerns are partly linked to the Remote Access Security Act (RASA), a proposed U.S. bill that would give the Commerce Department autonomy to regulate remote cloud-based access to restricted technology.
The House of Representatives passed the legislation in January by a vote of 369-22 and is awaiting action in the Senate. Chinese AI firms are cautious that the controversy surrounding Jensen Huang could garner extra momentum for the legislation. The altercation emerged after Wang said on Sunday that he had been unfaithful to his wife. The New York Times reported that the woman involved was Alice Wang, a former executive at Singapore-based cloud computing company Mega Speed, and that U.S. authorities had concluded she was a Chinese intelligence asset.
How Are Fears of Stern US Export Controls Affecting China’s Chip Market?
The subsequent effect of stringent enforcement is visible in the increased grey market activity in China. Traders and retailers have stockpiled these chips as they fear a broad U.S. crackdown on the unauthorized access of AI servers into China. A recent purchase by a Chinese university with deep ties to the People’s Liberation Army has highlighted the challenges facing Washington’s export-control efforts. According to procurement records cited by Reuters, the university bought Nvidia’s powerful H200 AI chips in October despite being included on a US export blacklist.
The purchase has raised further questions about how restricted processors continue to reach Chinese buyers. The issue spreads its roots to the movement of AI servers through third countries. In March, U.S. prosecutors charged Super Micro Computer co-founder Yih-Shyan “Wally” Liaw and two other individuals with conspiring to divert servers containing Nvidia chips to China through a Southeast Asia-based company. Prosecutors allege that the firm purchased approximately $2.5 billion worth of servers from Super Micro between 2024 and 2025, with a substantial portion allegedly diverted to China. The defendants have not been convicted.
Advanced Nvidia processors are essential for training large AI models used in chatbots and other generative AI applications. Any restrictions on overseas cloud access could make it arduous for Chinese developers to obtain high-end computing resources without directly importing the chips. As US lawmakers consider expanding export-control authority, Chinese AI firms are finding ways to gain computing capacity while bracing for the probability of stern constraints on both cloud access and physical chip supplies.
