Nvidia has said US export controls have effectively shut it out of China’s AI data-center market. Chinese documents tell a more complicated story.
A Reuters review published Nov. 13 found more than 23 billion yuan, or about $3.2 billion, in Chinese AI computing projects and equipment agreements that explicitly cite Nvidia’s restricted H100, H200 and H800 processors. The chips are banned, but demand for them evidently persists.
Where the chips are showing up
The records Reuters examined cover three categories: planned investments, awarded contracts and computing facilities already in operation. That range matters.
Plans may be aspirational. Awarded contracts and active facilities are considerably harder to dismiss.
The projects span China’s wealthy coastal technology hubs and remote western provinces. Reuters highlighted activity in Fujian, Shaanxi, Inner Mongolia and Gansu.
The largest example is a 6.2 billion yuan ($873 million) AI center in Fujian. Its requirements specifically name H100 processors, which became a workhorse of the global AI buildout.
In Shaanxi, a 3 billion yuan project linked to state-owned Sinomach involves H100 and H200 chips. The involvement of a state-owned enterprise in a project built around export-controlled American hardware is a detail likely to draw attention in Washington.
In Gansu, a facility was awarded a 1.16 billion yuan contract in July for 3,072 H100 and H200 processors.
The research does not establish whether, or how, the restricted chips have been delivered to each project. The documents do show that Chinese buyers are still designing computing infrastructure around hardware they are not supposed to be able to purchase.
Three years of controls, but demand persists
Washington began restricting exports of advanced AI chips to China in 2022. The rules have been tightened several times since, with each round intended to close gaps left by the previous one.
By some measures, Nvidia still has a substantial presence in the country. An IDC estimate cited in the research put the company’s share of Chinese AI-accelerator shipments at 34% in the first half of the year.
That figure sits uneasily alongside Nvidia’s public position. The company has cited regulatory uncertainty as the reason it excludes China data-center revenue from its forecasts.
What the findings mean for Nvidia, Washington and Beijing
For Nvidia, the most immediate concern is compliance exposure, rather than lost sales. The company has already excluded China from its data-center guidance, so its forecasts do not depend on these projects.
According to the research, calls in Congress for tighter controls are already growing. A paper trail exceeding $3 billion is the kind of evidence that can turn hearings into legislation.
Reuters identified growing US scrutiny as a risk for projects still trying to obtain banned Nvidia hardware.
Projects with chip-specific requirements built into their plans may face delays, redesigns or quiet reassessments if supply becomes even less reliable.
The broader takeaway from the documents is that export controls can restrict a product without curbing demand for it. More than 23 billion yuan in projects built around Nvidia’s restricted chips suggests Chinese buyers are not treating the ban as the final word.
