China's LineShine Tops Supercomputer Rankings by 20%, Ousting U.S. El Capitan
Updated
Updated · The New York Times · Jun 23
China's LineShine Tops Supercomputer Rankings by 20%, Ousting U.S. El Capitan
1 articles · Updated · The New York Times · Jun 23
Summary
LineShine was declared the world’s fastest supercomputer after standard benchmark tests showed the Shenzhen system outperforming El Capitan by more than 20%.
The result returns the supercomputing crown to China for the first time since 2017, intensifying a U.S.-China technology rivalry tied to science, national security and geopolitics.
Unlike most top-end machines, LineShine relies on standard microprocessors rather than GPUs, a design Top500 organizer Jack Dongarra said could better combine AI workloads with traditional scientific computing.
El Capitan at Lawrence Livermore National Laboratory had led the twice-yearly rankings since November 2024, making LineShine’s jump a notable shift in the global pecking order.
Is China's new supercomputer a brilliant innovation or a costly detour forced by US sanctions?
With unverified claims and major data breaches, how real is China's proclaimed technological supremacy?
Did US sanctions backfire, creating the self-sufficient Chinese tech rival they were meant to prevent?
LineShine: China’s 2.2 Exaflops All-CPU Supercomputer Challenges U.S. Dominance and Global HPC Benchmarks
Overview
On June 23, 2026, China unveiled the LineShine supercomputer at ISC26 in Hamburg, marking a major milestone in high-performance computing. LineShine achieved a verified HPL score of 2.198 exaflops, making it the world’s fastest supercomputer and the first all-CPU, fully domestic exascale system. Built on the LingKun platform, it uses powerful LX2 304-core processors and the proprietary LingQi interconnect, relying entirely on domestically produced components. This achievement highlights China’s return to global supercomputing leadership and demonstrates its ability to innovate independently in advanced technology.