Reuters has identified more than $3.2 billion worth of Chinese AI computing projects designed around Nvidia’s restricted processors — even though US export controls have, by Nvidia’s own account, shut it out of China’s data-centre market. Why it matters: the findings show Chinese demand for top-end Nvidia hardware has not disappeared; it has moved into local-government ventures, universities and regional computing centres that are now exposed if Washington tightens the rules further.
November 13, 2026
What Reuters found
Reuters reviewed more than 150 Chinese procurement notices, investment plans, construction contracts and corporate disclosures. Together they describe projects and equipment agreements with a combined reported value above 23 billion yuan ($3.2 billion), built around Nvidia’s restricted H100, H200 and H800 processors. The projects stretch from China’s wealthy coastal technology hubs to remote western provinces, and they reach well beyond the country’s largest tech firms into local government-backed ventures, universities and regional computing centres.
Two cautions are built into the number itself. The $3.2 billion total covers buildings, power systems and other infrastructure, not spending on Nvidia chips alone. And it mixes planned investments and signed contracts with installations that are already complete.
Demand that did not go away
Nvidia has said US export controls have effectively shut it out of China’s AI data-centre market. Jensen Huang, who resigned as chief executive on Thursday, said last year that Nvidia’s market share in China had fallen from 95% to zero, and the company has excluded China data-centre computing revenue from its forecasts, citing regulatory uncertainty. Its August filing said it was effectively excluded from that market.
The documents Reuters reviewed tell a different story about demand: Chinese customers are still planning purchases around restricted processors and, in some cases, operating them. One facility in northwestern Gansu province was expected to install thousands of Nvidia chips before year-end, though Reuters could not establish whether those processors ever arrived. An IDC report cited by Reuters in October estimated Nvidia still accounted for 34% of Chinese AI-accelerator shipments in the first half of 2026. Those figures measure different things, and Reuters’ reporting does not establish how the restricted hardware it found reached Chinese users.
Why these projects are now at risk
Washington began restricting Nvidia’s advanced AI processors in 2022 and has tightened the controls since, including on the H800, a modified chip developed for China. In January, the Trump administration moved to review H200 export licences for eligible Chinese customers individually, subject to conditions. Some of the facilities Reuters found are already operational or under construction, which leaves their operators exposed to delays, supply disruptions or costly redesigns around domestic chips if controls tighten again.
Huang’s departure could deepen that exposure by weakening Nvidia’s lobbying leverage in Washington, where lawmakers are pressing for tougher curbs. Sunny Cheung, an analyst at the Washington-based Jamestown Foundation, told Reuters: “Washington must stop focusing solely on who buys the chips on the surface and start examining who ultimately uses the computing power — and for what purpose.”
Reuters notes the projects it identified have not been linked to Megaspeed, the Singapore-based Nvidia cloud partner whose former executive is at the centre of the allegations surrounding Huang, and that there has been no public finding that Huang violated export-control laws. Megaspeed has denied illegally diverting Nvidia chips to China, and Nvidia’s board has said its independent investigation found no reason to believe company secrets or US national security were compromised. China’s commerce ministry and Nvidia did not immediately respond to Reuters’ requests for comment.
Why it matters
For anyone buying, building or investing around AI compute, this is the gap between policy on paper and hardware on the ground. Export controls have cut Nvidia’s official China business to near zero, yet billions of dollars of Chinese infrastructure is still being designed around its best chips. If Washington now moves from restricting sales to tracking end users — as the proposals in Congress would — projects like these are where the enforcement fight lands first. That is the “everything AI, tested” takeaway: the control regime is only as strong as its visibility into who is actually using the compute.
Frequently asked questions
How much are the Chinese projects Reuters found worth?
More than 23 billion yuan, or about $3.2 billion, across 150-plus procurement notices, investment plans, construction contracts and corporate disclosures. The total includes buildings, power and other infrastructure, not just chips.
Which Nvidia chips are the projects built around?
Reuters found projects designed around the restricted H100, H200 and H800 processors. US controls on those chips began in 2022 and have been tightened since.
Does this prove Nvidia sold restricted chips into China?
No. Reuters’ figures do not establish how the restricted hardware reached Chinese users, and Nvidia says export controls have effectively shut it out of China’s data-centre market. The projects Reuters identified have not been linked to the Megaspeed allegations, and there has been no public finding that Jensen Huang violated export-control laws.
What happens to these projects if US controls tighten?
Operators face possible delays, supply disruptions or expensive redesigns around domestic Chinese chips, according to Reuters — a risk that may grow as US lawmakers push for tougher restrictions after Huang’s resignation.
Sources: Reuters