Chinese developers are making significant strides in the AI sector, planning to invest over $32 billion in data centers that heavily rely on NVIDIA processors. According to the results published in the material, this ambitious move comes despite the ongoing restrictions from the United States on advanced chip sales to China.
NVIDIA's Cutting-Edge Chips
According to a Reuters investigation, these projects are focused on NVIDIA's cutting-edge chips, including the following models:
- H100
- H200
- H800
Proposed Computing Centers
A notable example is a proposed computing center in Fujian province, which is estimated to cost around 62 billion yuan. Additionally, a state-owned engineering group in Shaanxi has secured a contract for a 3 billion yuan center that specifically calls for the use of H100 and H200 chips.
Impact of US Export Controls
Since 2022, the US has imposed restrictions on the sale of advanced AI chips to China, citing national security concerns. This has created a challenging environment for Chinese developers, who now face the dilemma of whether to continue investing in NVIDIA technology or explore alternative processors. The situation underscores the critical role NVIDIA plays in China's AI infrastructure plans and raises questions about the effectiveness of US export controls.
In a notable development, Alphabet has announced plans to sell its custom Tensor Processing Units (TPUs) to external data centers, a move that contrasts with the heavy reliance on NVIDIA processors by Chinese developers. For more details, see more.
