Chinese AI projects specify $3 billion in Nvidia chips despite US bans

What happened

Over $3 billion worth of Chinese AI data center projects have been designed around Nvidia processors that are restricted by US export controls. These projects include both planned and operational facilities across various provinces in China. Despite US regulations, Chinese organizations continue to specify Nvidia chips in their infrastructure plans. Nvidia CEO Jensen Huang resigned on November 12, 2026, amid a US-China intelligence controversy.

Key details

• More than $3.2 billion (23 billion yuan) in Chinese AI projects specify restricted Nvidia processors.
• Procurement notices and contracts cover over 150 projects, including both planned and completed facilities.
• Some projects, such as a 6.2 billion yuan center in Fujian and a 3 billion yuan project in Shaanxi, explicitly list Nvidia H100 and H200 chips.
• Operational clusters using restricted Nvidia hardware have been reported in Jiangsu and Fujian provinces.
• US export controls began in 2022 and have since tightened, affecting Nvidia's direct sales and market forecasts for China.

Why it matters

Chinese technology companies, local governments, and universities are affected because they continue to plan and sometimes operate AI facilities with Nvidia hardware, despite regulatory uncertainty and potential supply disruptions. US export controls create risks of delays, redesigns, and higher costs for these projects if access to Nvidia chips is further restricted. Nvidia has excluded China from its data center revenue forecasts, reflecting the impact of these controls on its business in the region.