China launches largest scientific AI computing cluster in Zhengzhou
China activated its largest scientific AI computing cluster on Tuesday at the National Supercomputing Internet's core node in Zhengzhou, Henan Province. The facility runs 60,000 domestically developed AI accelerator chips.
The cluster ranks among the world's top computing infrastructure by scale, comparable to systems operated by Meta and Google. Unlike those systems, all 60,000 chips are Chinese-made rather than built on Nvidia's architecture.
How researchers will use it
The cluster integrates datasets, tools, and more than 1,000 open-source large-language models. Researchers can request computing resources using natural language instead of writing code or configuring software.
Users describe their needs in plain terms. The system's "super scientific computing agent" automatically breaks down tasks, selects models, allocates resources, and returns results. This cuts the time needed to complete research work.
The infrastructure supports both scientific research and commercial AI development, though research applications drive its design.
Scale-up timeline
The core node began trial operations on February 5 with 30,000 AI accelerator chips. Tuesday's expansion to 60,000 chips marks the facility's full activation.
Chen Jing, vice president of the Technology and Strategy Research Institute, said the project demonstrates China's ability to deploy homegrown computing clusters at scale. It also signals progress in the country's push for self-reliance in chip manufacturing.
Policy context
China's 2026 Government Work Report included "coordination between computing and power" for the first time, calling for ultra-large intelligent computing clusters and coordinated computing-power development.
The Ministry of Industry and Information Technology announced Friday that it will support national high-tech zones in deploying intelligent computing facilities and high-quality datasets.
China's 15th Five-Year Plan (2026-30) outlines steps to advance the Digital China initiative. It calls for coordinated progress in computing infrastructure, model development, algorithms, and data resources.
For researchers working with large-scale computing, explore AI for Science & Research to understand how these systems apply to your work.
Your membership also unlocks:
