China plans a five‑year program to construct a nationwide network of AI data centers, with Bloomberg reporting an initial estimated expenditure of about 2 trillion yuan (roughly 91,197 billion forints). The authorities aim to link the facilities into a single IT network by 2028.
The initiative would be overseen by the National Development and Reform Commission, while operators would include telecommunications companies such as China Mobile and China Telecom.
Domestic chip supply targets
Part of the plan is to source at least 80 percent of the chips used — including AI accelerators — from domestic suppliers, among them Huawei. This is intended to reduce reliance on foreign technology amid long‑running U.S. restrictions on Chinese access to American technologies.
Industry estimates cited in reporting indicate that Chinese domestic suppliers could cover about 76 percent of China’s AI‑chip demand by 2030.
Technical and scaling challenges
Tom's Hardware cautions that the total project cost could exceed the initial 2 trillion yuan estimate once required upgrades to power and energy infrastructure are included; the publication suggests total spending might top 5 trillion yuan (about 228,004 billion forints).
Another constraint is manufacturing capability: Chinese production currently is largely at the 7‑nanometer node, which lags the most advanced Western manufacturing technologies. Scaling production is also a concern — for example, reports note that Huawei produced roughly 812,000 high‑bandwidth memory chips in 2025, a pace suppliers may struggle to maintain or expand.
Why this matters
If realized, the program would significantly bolster China’s capacity to support AI applications domestically and reduce dependence on foreign components. Analysts warn, however, that achieving the targets depends on advances in chip manufacturing, substantial energy‑sector investment, and the ability of suppliers to scale output; success could in turn diminish the effectiveness of U.S. sanctions.
