China has been expanding its computing infrastructure to empower a wide range of industries and support high-quality development, as all artificial intelligence applications depend on high-performance computing.
In the northeastern corner of Beijing, an AI computing power park has taken shape.
"Our core mission is threefold. First, to redefine AI computing infrastructure by building cost-efficient and plug-and-play AI factories that deliver stable intelligent computing services across industries. Second, to act as a bridge between AI and industry, accelerating the translation of AI capabilities into real productivity. Third, to nurture an open ecosystem by providing accessible computing power and practical application scenarios for our partners, thereby driving the growth of the core AI industry," said James Yang, chief marketing officer of Beijing Digital Economy Artificial Intelligence Data Center.
Inside the park complex, rows of domestically manufactured AI servers are neatly arranged. The site houses computing capacity measured in hundreds of peta-FLOPS (floating-point operations per second), equivalent to a million high-performance computers working simultaneously.
The facility also integrates full-stack AI capabilities, including computing resources, algorithms, models, along with AI tools. High-speed server clusters and real-time task allocation turn abstract notion of "computing power" into tangible, ready-to-use digital infrastructure.
"Traditional data centers are often closed black boxes -- people don't really know what's inside. We want to use simple, visual approaches to show the public how artificial intelligence can change their daily lives," said Yang.
By the end of 2025, China's total intelligent computing power surpassed 1,590 EFLOPS, ranking second globally.
EFLOPS, or ExaFLOPS, refers to a unit measuring computing performance, with each unit representing 1 quintillion FLOPS.
Computing power has become an integral part of everyday life and industry. In advanced manufacturing and research, industry models and intelligent agent platforms are driving transformation.
"It's easy to define computing power as a digital foundation -- like water or electricity. But that's not enough. In the AI era, the key to making it a true digital backbone is building unified systems for technology, service, and allocation to efficiently supply a large market," said Li Jingjing, director of the CSTCloud Department of Computer Network Information Center, Chinese Academy of Sciences.
China expands computing infrastructure to fuel high-quality development
