China boosts 'AI for Science' computing in 2 months without US chips
In early February, China started trials for its largest artificial intelligence (AI) computing cluster for scientific research, without using US chips. Now, in just two months, its AI computing infrastructure is fully operational, and the cluster’s domestically made AI accelerator chip count has doubled from 30,000 to 60,000, South China Morning Post reported.
The chips, developed by Sugon, a supercomputer developer affiliated with the Chinese Academy of Sciences, are running in the core node of the national supercomputing network in Zhengzhou. With this latest upgrade, the Zhengzhou core node is now China’s most powerful scientific intelligent computing infrastructure.
The development is “a breakthrough for China in computing infrastructure for AI-driven scientific research, which will help the country seize the commanding heights of AI industrial applications,” SCMP quoted Chinese state broadcaster CCTV.
Due to limited software access and a shortage of computing power, China has long struggled to conduct “AI for science” research, but the development of the Zhengzhou cluster is expected to help the country conduct critical research domestically and tackle complex scientific problems.
This expansion comes at a time when US lawmakers last week proposed a bill to further restrict China’s access. As of now, US laws severely restrict China’s access to advanced AI chips and manufacturing equipment (semiconductors) in a bid to prevent military advancement and protect technological dominance. Now, lawmakers have introduced a MATCH (Multilateral Alignment of Technology Controls in Hardware) Bill to close “critical gaps” in existing restrictions and bring allies such as the Netherlands and Japan into closer alignment on export controls within 150 days.
What is the US-China chip war all about?
In 2022, former US President Joe Biden signed a bill to boost domestic chipmaking, choking China’s tech ambitions by rolling out curbs on semiconductors. By passing the bill, the Biden administration aimed to limit China’s ability to develop technologies that could boost its military and financial strength. This came at a time of high-stakes tensions between the two global superpowers, already locked in a trade war and intensifying technology rivalry.
Later that year, Washington also restricted exports of some high-end microchips to China, citing national security concerns. This was followed by the US blacklisting 36 Chinese tech firms, restricting them from using American chip designs. The ban also restricted Nvidia and TSMC from selling their chips to Chinese firms.
Since then, the restrictions have further widened to include advanced chipmaking tools, despite resistance from US companies. Washington is now also pushing its allies to align with these controls.
China, in response, has accelerated efforts toward self-reliance by offering large subsidies, tax breaks and other incentives, while investing billions of dollars in domestic chip production, AI hardware and research. Companies such as Huawei and SMIC have emerged as symbols of Beijing’s push to reduce dependence on Western technology.
