Byagenze bite
The South China Morning Post reports that China’s Ministry of Industry and Information Technology has issued a five-year industry plan targeting 9,800 eflops of intelligent computing capacity by 2030. The plan also calls for 3.8 trillion yuan in cumulative information-infrastructure investment from 2026 through 2030 and the deployment of large computing clusters adapted to domestic chips.
The South China Morning Post reports that China’s Ministry of Industry and Information Technology included a target of 9,800 eflops of intelligent computing capacity in a five-year industry plan released on Monday. The plan also calls for 3.8 trillion yuan, or about US$532 billion, in cumulative information-infrastructure investment during 2026-2030.
According to the report, the plan calls for the orderly deployment of intelligent-computing clusters containing 10,000 graphics-processing cards, as well as clusters using 100,000 or more cards. It also calls for -computing facilities designed for different applications and for infrastructure to be adapted to domestically produced computing chips.
The SCMP reports that China’s intelligent-computing capacity reached 2,185 eflops at the end of June, up 177 per cent from a year earlier. The National Data Administration reportedly put the figure at about 2,450 eflops by the end of July. MIIT also reported that 52 intelligent-computing facilities had been built with more than 10,000 accelerator cards each.
The expansion builds on China’s East Data, West Computing project, launched in 2022 to move power-intensive workloads from densely populated eastern regions toward western areas with cheaper land and more abundant energy. The report says China has organized a network around eight national computing hubs, 10 data-center clusters and three regions coordinating computing facilities with power supplies.
Ibisobanuro birambuye: scmp.com ↗
Impamvu ari ngombwa
China’s target would require more than quadrupling the intelligent-computing capacity reported for June 2026, making AI infrastructure a central element of national industrial policy. The plan could influence where data centers are built, which chips and systems are purchased, and how quickly Chinese AI developers can obtain . The report does not independently confirm the plan’s funding, implementation timetable, or whether the targets will be met.
The reported plan makes capacity a measurable national objective rather than only a company-level investment decision. More available infrastructure could support training and for Chinese AI systems, although the report does not establish how much of the planned capacity will be usable by commercial developers or what proportion will support training versus inference.
The scale also highlights the practical constraints behind AI expansion. Building large clusters requires accelerator cards, electricity, networking, cooling and suitable data-center sites. China’s geographic redistribution strategy may reduce land and energy costs in some regions, but the report does not provide evidence that these constraints have been resolved.
The plan’s emphasis on home-grown chips is significant because it links AI capacity growth with domestic hardware development. The source does not independently test the performance, supply, compatibility or reliability of those chips, and it does not say whether the investment figure includes only public spending or broader information-infrastructure expenditure.
Uburyo bukoreshwa: Uburyo bukora
Shakisha ikoranabuhanga ryihishe inyuma yiri terambere.
Which component of an AI application is the machine-learning model itself?
Ibyo kureba
Key unknowns include how the planned investment will be allocated, which facilities will receive approval, how domestic chips will perform at the proposed scale, and whether power and network infrastructure can keep pace. The report describes national targets rather than a user-facing product, so there is no documented public access route or pricing.
Watch for the publication of implementation rules identifying responsible agencies, financing sources, construction milestones and criteria for approving 10,000-card and 100,000-card clusters.
Watch whether the reported capacity figures translate into accessible for AI companies, universities and other users, rather than remaining concentrated in state-backed or large corporate facilities. The source gives no access terms or pricing.
Watch for evidence on the performance and deployment scale of Chinese accelerators, as well as progress connecting western data centers to reliable power and high-capacity networks.
The report provides targets and ministry-attributed figures, but no independent audit of the capacity numbers or confirmation that the 2030 goal is financially and technically achievable.