China has launched a new supercomputing system called the Tianhe Xingyi, unveiled on December 6th by the National Supercomputer Center in Guangzhou. According to state media reports, the Tianhe Xingyi significantly outperforms China's previous top supercomputer, the Tianhe-2, in key metrics including processing power, networking capabilities, storage, and versatility for applications.
While specific technical details remain scarce, the new system reportedly uses advanced domestic chips and architecture. This allows China to continue advancing its supercomputing despite restrictions on the use of American processors in recent years.
National pride and technological self-sufficiency have driven China’s supercomputing pushes. The Tianhe-2 held the title of world’s fastest supercomputer from 2013 to 2016, putting China in the elite ranks globally. However, American export controls on high-tech chips forced China to innovate domestically.
The new Tianhe Xingyi supercomputer has “doubled in many aspects” compared to the Tianhe-2, according to Lu Yutong, director of the National Supercomputer Center in Guangzhou. This includes major performance gains in computing power, networking speed, storage capacity, and ability to run key applications.
The supercomputer will support needs across high-performance computing, AI model training, and big data analysis. This versatility suits Guangzhou’s growing technology hub, now ranking highly globally in innovation indexes. Guangzhou has invested heavily in R&D in recent years as well.
China has grown increasingly secretive about the specifications of its supercomputers. While the Tianhe-2 once topped global supercomputer rankings, China no longer discloses benchmark performance or participates in lists like the Top500. The country instead seems focused on domestic needs rather than proving its capabilities on the global stage. Still, the rapid development of systems like the Tianhe Xingyi makes clear that China aims to lead in supercomputing advancements.