NVDA•NVDA•China is drafting a 2 trillion yuan ($295 billion) plan over five years to build a nationwide network of interconnected AI data centers operated by state firms and using at least 80% domestic chip technology. This blueprint could sharply limit Nvidia’s role in China’s AI infrastructure even as U.S.-approved H200 chip exports remain delayed.
China’s National Development and Reform Commission is drafting a five-year plan to invest roughly 2 trillion yuan in a unified network of AI data centers. The goal is to interconnect fragmented regional computing hubs by 2028, delivering high-performance computing access across health care, finance, transportation and manufacturing sectors.
State-owned operators such as China Mobile and China Telecom are expected to manage the data centers, while local suppliers like Huawei Technologies will provide at least 80% of the core AI chips and networking equipment. Nine homegrown AI chips recently passed security review, positioning domestic players for strategic deployments.
Although Washington has approved sales of Nvidia’s previous-generation H200 chips, shipments have yet to begin, suggesting Beijing’s focus on domestic capacity could restrict Nvidia’s market share. Advanced Micro Devices may face similar constraints as China pushes to replace foreign hardware with homegrown solutions.
Funding is expected from sovereign debt—ultra-long-term government bonds—and state industry funds, with possible bank loans and private capital supplements. If power grid integration is included, total investment could reach at least 5 trillion yuan, signaling a broader infrastructure cycle beyond data centers.
Barrons
