China’s AI industry is setting its sights on a massive leap forward with plans to install over 100,000 NVIDIA AI chips in hyperscale data centers across the nation. Despite the challenges posed by international trade restrictions, this ambitious project seems determined to push forward. The AI facilities, planned for Yiwu in eastern China, are part of an extensive rollout that could see more than 115,000 high-end NVIDIA chips deployed across 36 data centers, mostly in China’s western deserts. 
However, this project raises an important question: how will China get access to these vital chips?
The United States has restricted the export of AI chips to China, creating a significant roadblock for the country’s AI ambitions. While the government is reportedly working with regional allies like Malaysia and Singapore to access the technology, the scale of the project suggests that Beijing is pursuing alternative methods to sidestep these restrictions. Whether through trade loopholes or local solutions, China seems determined to get its hands on the necessary hardware, despite the political and legal challenges.
Interestingly, sources in the U.S. seem unaware of the scale of the project, which could mean that it is being developed under a veil of secrecy. If China pulls this off, it would mark one of the largest AI infrastructure builds in the country’s history, though it’s unlikely to rival the scale of the U.S.’s AI infrastructure. Even so, the move shows that China is working relentlessly to establish itself as a global leader in AI, even if it means circumventing international barriers.
The alternative to NVIDIA’s chips in China could be domestically developed options like Huawei’s Ascend, which some claim can rival NVIDIA’s Hopper and Blackwell chips. Yet, despite these claims, it seems that China’s reliance on NVIDIA’s AI chips remains strong for now. With its growing demand for advanced computing power, China’s AI sector could become a key player in the global tech race.