Z.AI Builds 1-Gigawatt Data Center on Chinese Chips

Beijing-based Z.AI, formerly known as Zhipu AI and spun out of Tsinghua University, is one of China’s leading developers of large language models. Its push into domestic computing infrastructure has gained strategic weight as U.S. export controls restrict Chinese access to Nvidia’s most advanced accelerators. Building large training systems around homegrown processors would give Chinese AI companies a more durable supply chain, though domestic chips can still lag Nvidia hardware in performance and energy efficiency.

Bloomberg reported on July 20, 2026, that Z.AI had completed a data center designed to draw about 1 gigawatt of power and run exclusively on Chinese-made chips, with part of the facility already operating. The company has also built or operates several computing clusters containing more than 10,000 chips each. Z.AI did not disclose the project’s investment cost or identify all chip suppliers, but the scale offers a significant test of whether China can train GLM models without relying on restricted Nvidia hardware.