China plans to invest about CN 2 trillion (~$295 billion) in building a nationwide network of interconnected data centers over the next five years. According to Bloomberg, the project will be overseen by key government agencies led by the National Development and Reform Commission (NDRC), while infrastructure and communications channels will be managed by state-owned operators China Mobile and China Telecom. The main goal of the initiative is to advance the AI sector and strengthen China’s technological sovereignty in the face of the US.

A key feature of the project will be to rely on domestic equipment: about 80% of the infrastructure will be built with Huawei and other Chinese manufacturers' chips, effectively closing the project to Nvidia and AMD. Nvidia CEO Jensen Huang publicly acknowledged back in May 2026 that the company had "largely succumbed" to the Chinese market for advanced AI accelerators, and government entities and large businesses were consistently switching to Cambricon solutions, Biren and Huawei as part of the Xinchuang policy.

The main beneficiary of the program is Huawei with its Ascend chip line (910C, 950PR and new developments): the company forecasts AI revenue at around $12 billion in 2026 (+60% year-on-year). The largest customer was ByteDance, with orders exceeding $5.6 billion this year alone, among other customers being Alibaba and Tencent. In parallel, Huawei is developing the CANN software stack as an alternative to Nvidia’s CUDA and expanding its Atlas cluster lineup, which will allow China to build one of the world’s largest centralized computing infrastructures based solely on its own component base.