China’s $3 billion-plus pipeline of Nvidia-powered AI data centres

A Reuters investigation has revealed a $3 billion network of Chinese data centers powered by Nvidia's restricted processors, despite Nvidia's claim that US export controls have effectively barred it from the Chinese AI data center market. The analysis reveals projects spread across China's major technology centers and remote regions, demonstrating how deeply Nvidia's technology is embedded in China's AI expansion.

The findings follow Nvidia co-founder Jensen Huang's resignation as CEO following a scandal, which could exacerbate vulnerabilities for Chinese companies facing potential delays and supply disruptions as US restrictions tighten.

Several projects identified by the Reuters investigation include a proposed 6.2 billion yuan ($873 million) AI computing center in coastal Fujian province, a 3 billion yuan computing center in northwestern Shaanxi province, and a 2.26 billion yuan project in northeastern Jilin province that combines Nvidia H100 processors with Chinese-developed chips. In Inner Mongolia, a 1.5 billion-yuan development plans to install 625 Nvidia H200 servers, with construction already underway on two buildings in its first phase.

These plans highlight developers' reliance on US processors, which are vulnerable to trade restrictions. Tighter curbs risk construction delays, increased costs, or forced redesigns around domestic alternatives. In Gansu province, a local company was awarded a 1.16 billion yuan construction contract in July for a computing center that plans to install approximately 3,072 Nvidia processors. Other Chinese facilities, such as a computing cluster using Nvidia H800 processors in Jiangsu province, are already in operation.

Smaller purchases demonstrate demand extending to research institutions, with Shenzhen University awarding a contract for three H100 processors and Northwestern Polytechnical University buying two H200 chips.