China has launched its largest artificial intelligence computing cluster dedicated to scientific research, marking a significant leap in its push for technological self-reliance. The system, which entered full operation on Tuesday, has doubled its number of domestically produced AI accelerator chips to 60,000 in just two months, according to state broadcaster reports. The infrastructure is anchored at a core node in Zhengzhou, in Henan province, and is now considered the most powerful scientific intelligent computing platform in the country.
The AI acceleration cards powering the cluster were developed by Sugon, a leading Chinese supercomputer firm affiliated with the Chinese Academy of Sciences. The rapid expansion from 30,000 chips during trial operations in February highlights the pace at which China is scaling its high-performance computing capabilities. Officials describe the development as a breakthrough in building advanced infrastructure for “AI for science,” a field that applies artificial intelligence to complex scientific discovery.
For years, Chinese researchers have faced constraints in this area due to limited computing power, software bottlenecks, and reliance on foreign technologies. The new system seeks to eliminate those barriers. Built entirely with domestic technology—from chips to interconnects and software platforms—the cluster enables researchers to run sophisticated simulations and process high-dimensional scientific problems without depending on external suppliers.
The platform is designed to significantly accelerate research timelines across multiple disciplines. In drug development, scientists have achieved dramatic gains, with protein folding simulations running thousands to millions of times faster than before. Tasks that once required years can now be completed in a matter of days. Similar progress has been reported in materials science, where screening processes have been reduced from years to days.
Beyond life sciences and materials research, the computing cluster supports large-scale simulations in fields such as brain science and aerospace engineering. It has already demonstrated the ability to simulate tens of billions of atoms, model the human brain’s 86 billion neurons, and conduct turbulence simulations at unprecedented scales involving trillions of computational grids.
Complementing the hardware, Sugon has introduced a one-stop development platform called OneScience, which hosts a wide range of scientific models and datasets. The platform is designed to be accessible even to researchers without programming expertise, lowering the barrier to entry for advanced computational research and expanding the pool of potential users.
The launch comes amid intensifying global competition over advanced technologies. In the United States, lawmakers are pushing for stricter export controls on semiconductor manufacturing equipment, aiming to further limit China’s access to cutting-edge tools. At the same time, Washington has introduced initiatives to build its own integrated AI platforms for scientific research, signaling a deepening rivalry in the race to harness artificial intelligence for innovation.
Against this backdrop, China’s new AI supercluster represents both a technological milestone and a strategic move to secure independence in critical research infrastructure. By reducing reliance on foreign components and dramatically increasing computational capacity, the country is positioning itself to accelerate scientific discovery and compete more aggressively on the global stage.
