China commits $532B to quadruple AI computing capacity by 2030, built entirely on domestic chips
A five-year plan from China's industry ministry targets 9,800 exaflops of compute -- more than four times current capacity -- deliberately avoiding reliance on foreign chips given US export controls.
China's Ministry of Industry and Information Technology published its 15th Five-Year Plan for the information and communications sector on September 7, 2026, committing 3.8 trillion yuan (roughly $532 billion) in cumulative infrastructure investment through 2030. The target: 9,800 exaflops of intelligent computing capacity, more than four times the 2,185 exaflops reported as of end-June 2026 -- itself already up 177% year over year. The plan is explicit that this capacity will be built on domestic chips, a deliberate response to US export controls rather than a cost decision.
Scale as policy, not incidental growth
The plan calls for "orderly deployment" of compute clusters with 100,000 or more accelerator cards each; China already operates 52 facilities at that scale. Framing a national compute target in exaflops, with an explicit domestic-silicon requirement attached, is industrial policy aimed squarely at chip self-sufficiency, not a general technology-investment announcement.
Why this matters for anyone modeling global AI compute supply
A government committing over half a trillion dollars to quadruple compute capacity on domestically fabricated chips is a structural signal, not a one-off data center announcement -- it changes the medium-term picture for global GPU and AI-chip supply/demand, and for how exposed any company's supply chain is to US-China export-control policy shifts. For finance and infrastructure leaders modeling where AI compute capacity comes online next, sovereign buildouts like this one are increasingly as relevant as any single hyperscaler's capex guidance.