Analysis
China's Ministry of Industry and Information Technology published its 15th Five-Year Plan for the information and communications sector on September 7, setting a target of 9,800 eflops of "intelligent computing" capacity by 2030 -- more than four times the roughly 2,185 eflops China had reached by the end of June, which was itself up 177% from a year earlier. The plan calls for 3.8 trillion yuan (roughly $530 billion) in cumulative information-infrastructure investment over the 2026-2030 period, per Xinhua.
The numbers matter more as a policy signal than as a precise forecast: China is publishing a specific, trackable national compute target the way it has historically set steel or semiconductor output targets, which gives state-linked cloud providers and chipmakers a planning number to build capital expenditure around. The plan calls specifically for "orderly deployment" of computing clusters using 10,000 or 100,000-plus graphics processing cards, and for adapting infrastructure to run on domestically designed chips -- a direct response to continued US export restrictions on advanced Nvidia and AMD accelerators, and a natural extension of the chip-independence push Pulse covered this week in Huawei's own Kirin 9050 Pro launch.
For US labs and chipmakers, the practical read is less about whether China hits 9,800 eflops precisely by 2030 and more about the trajectory: quadrupling compute in five years, even short of the target, keeps pressure on the assumption that export controls meaningfully cap China's AI capacity rather than just slowing it and pushing it toward domestic silicon.
