Despite claims from Nvidia that U.S. export controls have effectively barred the company from China’s artificial intelligence data-centre market, an investigation has identified over $3 billion worth of Chinese computing projects built around the chipmaker's restricted processors. The findings, based on an analysis of more than 150 procurement notices, investment plans, and construction contracts, reveal that demand for Nvidia hardware persists across China, from major coastal tech hubs to remote western provinces.
The reported $3.2 billion figure encompasses planned investments, signed contracts, and completed installations, including infrastructure such as power systems and facilities. These projects involve a wide range of users, including universities, regional computing centres, and ventures supported by local governments. Documents reviewed show that Chinese customers continue to plan purchases and, in several instances, operate restricted hardware like the H100, H200, and H800 processors.
The scrutiny of these projects follows the recent resignation of Nvidia co-founder Jensen Huang, who stepped down as chief executive amid allegations regarding his relationship with a businesswoman suspected by U.S. authorities of having intelligence ties to Beijing. While Nvidia’s board has launched an independent investigation, the company stated it has no reason to believe national security or company secrets were compromised. There has been no public finding that Jensen Huang violated export-control laws.
Washington began restricting advanced Nvidia AI processors in 2022, subsequently tightening controls on modified versions developed specifically for China. In January, the Trump administration mandated that H200 export licences for Chinese customers be reviewed on an individual basis. Despite these regulations, an October IDC report cited by Reuters noted that Nvidia accounted for 34% of Chinese AI-accelerator shipments in the first half of 2026.
Key facts
- Over 150 procurement documents confirm more than $3 billion in Chinese projects utilize restricted Nvidia technology.
- Projects include hardware installations in regional computing hubs and commercial sites in provinces like Gansu.
- An IDC report estimated Nvidia held 34% of Chinese AI-accelerator shipments in the first half of 2026.
- Jensen Huang resigned as Nvidia CEO on Thursday following a scandal involving a former cloud partner executive.
- U.S. export controls currently restrict the sale of advanced processors like the H100, H200, and H800 to China.
Why it matters
The continued reliance on restricted Nvidia hardware highlights the challenges Washington faces in curbing Beijing’s access to technologies fueling the global AI boom. Analysts suggest that current enforcement efforts may need to expand beyond initial chip purchases to track the final users and the specific purposes of the computing power.
Background
Nvidia has previously stated that its market share in China fell from 95% to zero due to U.S. regulatory intervention. The company has excluded China data-centre computing revenue from its recent financial forecasts.
Source: Economic Times
Comments