Beijing is making the most coordinated effort to date for the nationwide development of artificial intelligence (AI) infrastructure, aiming to surpass the US in a technology that could change the global landscape.
China plans to spend approximately $295 billion over the next five years to build data centers across the country. As part of this, government agencies such as the National Development and Reform Commission are drafting a plan to create a network of interconnected computing hubs nationwide, to be managed by state-owned companies including China Mobile and China Telecom, among others.
The idea is to rely on domestic suppliers, such as Huawei Technologies, for at least 80% of the technology, including AI chips, effectively excluding US giants like Nvidia and Advanced Micro Devices, according to Bloomberg.
This overall plan echoes earlier initiatives that mobilized resources to support national champions like Huawei, with the aim of replacing US technology. At the same time, it is a pillar of the "Six Networks" program announced earlier this year, which involves building critical infrastructure from water and electricity to computing networks. Analysts note that Beijing is determined to invest in cutting-edge technologies even as spending in other areas is cut due to rising government debt.
Most of the funding will come from government borrowing, including special long-term government bonds with maturities over ten years, as well as state investment funds supporting strategic industries. The amount will be supplemented by bank loans and private capital, according to well-informed sources. According to Charlie Dai, principal analyst at Forrester Research, a unified computing network will consolidate fragmented regional resources and provide businesses with broader access to high-performance computing power. It will also accelerate the evolution of AI models and the development of agentic and physical AI services across many sectors of the Chinese economy.
The idea of building a nationwide computing network was presented in China's latest five-year plan, covering the period until 2030, in which Beijing pledged to prioritize the construction of data centers. However, the latest investment target—which had not been previously publicized—remains smaller compared, for instance, to the $725 billion that US giants like Meta and Microsoft plan to spend on AI this year alone.
Chinese data centers, however, generally cost less than those in the US, due to lower labor, component, and construction costs, as well as incentives provided by local governments.
It remains unclear how the planned unified data center network will operate alongside private facilities. The main goal is to connect these scattered infrastructures into a cohesive network by 2028. If the plan ultimately goes ahead, Chinese companies are expected to be the main beneficiaries.
The US has allowed Nvidia to sell its previous-generation H200 AI chips to the Chinese market, a significant easing of restrictions imposed to curb AI development in China. However, shipments of the components have not yet started, indicating that Beijing is increasingly confident about replacing part of its AI computing power with domestically made equipment. In May, nine types of Chinese AI chips—including products from Huawei, Alibaba, and Moore Threads Technology—successfully passed a security review by a Chinese technology security agency, paving the way for their wider use.
"Every player in the ecosystem will benefit," concludes Charlie Dai of Forrester Research.
