China's Intelligent Compute Scale Expands 2.8-Fold as Domestic AI Supercluster Enters Service

According to China's National Development and Reform Commission (NDRC), the country's intelligent computing capacity by the end of June reached 2.8 times its scale from the same period last year. Speaking at a press conference reported by People's Daily, NDRC spokesperson Jiang Yi highlighted key technological milestones, including the deployment of China's first fully domestic 100,000-card artificial intelligence supercluster and global downloads of domestic AI models exceeding 10 billion. Domestic AI-related industries maintained a growth rate above 30 percent over the first half of the year.

The 2.8-fold expansion in intelligent computing capacity and the operational launch of a fully domestic 100,000-card supercluster signal that China is making progress in bridging software and hardware ecosystems for domestic chips and models. This infrastructure expansion aims to strengthen digital economy supply chains and foster new quality productive forces amid global competition.

China's intelligent computing capacity reached 2.8 times its scale compared to the same period last year as of the end of June, according to official details released by the National Development and Reform Commission (NDRC) and reported by People's Daily. Speaking at the commission's monthly press conference on July 31, Jiang Yi, spokesperson and director of the Policy Research Office at the NDRC, stated that China's independent innovation in artificial intelligence accelerated notably during the first half of the year, achieving rapid breakthroughs across multiple links in the full industry chain. A major focus of the NDRC's announcement was the reinforcement of China's computing base. In particular, the country's first fully domestic 100,000-card artificial intelligence supercluster has been officially put into operation. The rollout of this ultra-large supercluster underscores ongoing national efforts to construct high-capacity compute infrastructure grounded in domestic hardware components, strengthening the foundational infrastructure required for heavy AI workloads. In addition to expanding hardware capacity, model capability has advanced significantly. Domestic enterprises such as DeepSeek and Moonshot AI have successively released multiple open-source large models featuring parameter sizes at the trillion level. The NDRC reported that cumulative global downloads of Chinese domestic models have now surpassed 10 billion, while technical adaptation between domestic model software and indigenous compute chips has been accelerating across the industry. The flow of core data elements has also gained momentum, with more than 120,000 high-quality datasets constructed to date. According to Jiang, a positive cycle wherein computing power, model iteration, and data resources mutually enhance one another is rapidly forming. Driven by these coordinated developments across compute, models, and data, artificial intelligence has emerged as a major dynamic driver of broader economic expansion, with related AI sectors continuing to sustain high growth rates above 30 percent.