Chinese artificial intelligence companies are seeking to secure access to Nvidia-powered data centers outside China, fearing that a controversy involving the chipmaker’s chief executive, Jensen Huang, could give Washington fresh impetus to restrict their ability to use advanced US processors remotely.
Three people familiar with the matter told Reuters that a major Chinese cloud provider had discussed internally whether to lock in overseas Nvidia computing capacity, while other companies had approached data-center operators about reserving access. Smaller AI developers are also considering similar arrangements, including contacting major computing capacity providers in Southeast Asia.
The activity points to growing concern among Chinese technology companies that Washington could close a gap in its export control regime. Although the United States restricts exports of Nvidia’s most advanced artificial intelligence processors to China, companies can still access some of the same computing resources by renting capacity at data centers in other countries.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
That access has become an important route for Chinese developers seeking the processing power required to train and operate large AI models without importing restricted chips directly.
The companies are concerned that the allegations surrounding Huang could strengthen support for legislation that would allow the US government to regulate such arrangements, the people said. They declined to identify the cloud provider or their companies because of the sensitivity of the discussions.
At the center of the concern is the Remote Access Security Act, or RASA, a proposed US law that would give the Commerce Department authority to regulate remote, cloud-based access to restricted technology.
The House of Representatives passed the bill in January by 369 votes to 22, and it is awaiting action in the Senate. If enacted, the legislation could give Washington a more direct means of controlling how foreign companies access advanced computing resources hosted outside the United States.
Current export restrictions limit the direct sale of Nvidia’s most advanced processors to China, but overseas data centers can provide access to similar computing power through cloud services. Extending controls to remote access would make that route more difficult and potentially increase the cost and complexity of developing advanced AI systems.
Nvidia processors are widely used to train and operate the large language models behind chatbots and other generative AI applications. For Chinese companies competing in the global AI market, access to high-performance computing is therefore a commercial and technological necessity.
The latest scramble suggests companies are considering securing capacity before any new restrictions take effect. Reserving overseas computing resources could provide a degree of continuity if Washington expands its controls, although such arrangements would not necessarily protect customers from future regulatory action.
Huang’s personal controversy has added political sensitivity to an already contentious export control debate. He said on Sunday that he had been unfaithful to his wife. The New York Times reported that the woman involved was Alice Huang, a former executive at Singapore-based cloud computing company Megaspeed, and that US authorities had concluded she was a Chinese intelligence asset.
According to two people familiar with the matter, Alice Huang had business and investment ties with several major Chinese technology companies, including Tencent and Alibaba. Both companies did not immediately respond to requests for comment.
Singapore business filings reviewed by Reuters show that Alibaba subsidiary Apex Enterprise Solutions provided financing to Speedmatrix, a Malaysian company linked to Megaspeed.
Megaspeed has faced scrutiny over suspicions that it helped Chinese customers obtain advanced Nvidia processors subject to US export controls. The company has denied illegally diverting chips to China. Nvidia, Megaspeed and the US Commerce Department did not immediately respond to requests for comment.
The allegations have created a politically sensitive situation for Nvidia, whose products sit at the center of the global AI infrastructure market. However, the immediate concern for Chinese customers is less about the company’s internal leadership than the possibility that Washington could use the controversy to accelerate restrictions on access to its technology.
Grey-Market Chip Demand Rises As Enforcement Risks Increase
The uncertainty is also affecting China’s market for restricted Nvidia processors. Two of the people said the prospect of tougher US enforcement had encouraged traders and intermediaries to stockpile advanced AI chips in anticipation of a broader crackdown on smuggling.
That activity points to a second channel through which tighter controls could affect Chinese AI development. Companies unable to obtain sufficient computing capacity through overseas cloud services may turn to intermediaries offering restricted processors through unofficial supply chains, potentially exposing buyers to greater legal and supply risks.
Recent cases have highlighted the difficulty of preventing advanced chips and servers from reaching China through third countries.
Procurement records show that a Chinese university with deep ties to the People’s Liberation Army bought Nvidia’s powerful H200 AI processors last month, even though the institution is on a US export blacklist. The purchase highlights gaps in Washington’s efforts to prevent restricted technology from reaching Chinese institutions.
In March, US prosecutors charged Super Micro Computer co-founder Yih-Shyan “Wally” Liaw and two other people with conspiring to divert servers containing Nvidia chips to China through a company based in Southeast Asia.
Prosecutors alleged that the company bought about $2.5 billion worth of servers from Super Micro between 2024 and 2025, with a substantial portion subsequently diverted to China. The defendants have not been convicted.
The case illustrates how export restrictions can be circumvented through complex procurement and distribution networks, even as US authorities increase scrutiny of suppliers and intermediaries.
For Chinese AI companies, the emerging risks extend beyond the direct availability of Nvidia processors. A tougher regulatory framework covering overseas cloud access could narrow one of the remaining routes to advanced computing, while stronger enforcement against smuggling could make unofficial procurement more difficult.
The combination could increase pressure on Chinese developers to secure computing resources in advance, rely more heavily on domestic chip suppliers, or find ways to make existing hardware more efficient.