China Expands AI Computing Scale In Two Months, Without US Chips
This image was generated by AI and may not depict real events.
China has expanded its AI computing infrastructure to 60,000 accelerator cards at a national supercomputing node in Zhengzhou, without using US chips. The development is seen as a strategic move to build sovereign AI infrastructure amid US export controls.
China's national supercomputing network in Zhengzhou has reached full operation with 60,000 AI accelerator cards, doubling its capacity in two months. The cards were produced by Sugon, a Chinese supercomputer developer. This development is a response to US export controls restricting China's access to advanced processors. The Zhengzhou cluster is a national resource for scientific research, centralizing high-performance AI compute under domestic control. Chinese researchers have faced constraints, including computing shortages and reliance on foreign toolchains. The Sugon accelerator cards are part of a growing portfolio of Chinese-designed AI chips. Beijing is compensating for the performance gap with US chips through scale and integration, building large clusters and investing in software optimization.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.