Chinese Firms Seek to Reserve Nvidia Capacity Abroad as US Restrictions Fears Grow
Key Facts
Chinese AI companies are seeking to secure computing capacity using Nvidia chips at data centers outside China, amid fears that attention surrounding Jensen Huang could prompt tighter US restrictions on remote access. Sources familiar with the matter told Reuters that a major Chinese cloud provider had internally discussed installing overseas computing capacity, while other companies had contacted data center operators to ask about reserving it. Smaller labs are also considering turning to rental computing providers in Southeast Asia. These arrangements allow processors to be used remotely by renting capacity at data centers outside China, despite US restrictions on exports of Nvidia’s latest chips to the Chinese market. The moves reveal an effort to secure access to advanced computing resources through services across borders, while rules governing such use remain under discussion.
The concern is linked to the Remote Access Security Act, which could give the US Commerce Department authority to regulate cloud access to restricted technology. The House of Representatives passed the bill on January 12, 2026, by 369 votes to 22, while 39 members did not vote. Companies fear that the attention surrounding Huang could give the bill political momentum, broadening the debate from chip shipments to the use of computing capacity available through overseas data centers. That ties overseas rental arrangements to a legislative process that could affect how restricted technology is accessed. The available evidence did not indicate that new rules specifically governing remote access had taken effect.
Fears also extended, according to sources cited by Reuters, to the market for restricted chips inside China: the prospect of tighter enforcement fueled demand in the gray market, and traders and brokers began stockpiling the processors. Procurement records showed, according to the report, that a Chinese university with close ties to the People’s Liberation Army bought H200 chips in the previous month, despite being on a US export blacklist. In a separate case, US prosecutors said a company linked to the matter bought nearly $2.5 billion worth of servers from Super Micro between 2024 and 2025, a large portion of which was diverted to China. The defendants in the case have not been convicted. Singapore business records reviewed by Reuters also showed that an Alibaba subsidiary financed Malaysia’s Speedmatrix, which is linked to Megaspeed.
Megaspeed denied illegally diverting chips to China. Reuters said it could not independently verify allegations that Alice Huang was linked to Chinese intelligence, which China denied. These developments come as licensed shipments of H200 chips to China remain limited, according to Nvidia disclosures. The company said the US government had granted licenses to ship small quantities of H200 products to specific customers, but it had been unable to sell all the licensed products because of Chinese government restrictions. Shipments made under the licensing program accounted for less than 1% of data center revenue in the latest reported quarter. Nvidia also said it assumed no revenue from data center computing in China in its forecast for the third quarter of its fiscal 2027.
Nvidia recorded an expense of $0.4 billion related to excess H200 inventory and purchase commitments during the first half of its fiscal 2027, and said demand for these products had declined. In the second quarter of the same fiscal year, the company’s revenue was $96.221 billion, up 18% from the previous quarter and 106% from the same period a year earlier. The figures provide context for the scale of the company’s business, at a time when licensed H200 shipments to China do not represent a large share of data center revenue. Nvidia shares closed at $166.19 on November 12, 2026. The company is scheduled to report third-quarter results on November 17, 2026, for the period ended October 25, 2026.