China’s LineShine returns country to top of global supercomputing with 2.198 exaflops

China’s domestically developed LineShine supercomputer has become the world’s fastest system, reclaiming the top spot in global supercomputing for the first time in nine years. Announced at the ISC 2026 conference in Hamburg, Germany, LineShine recorded a sustained performance of 2.198 exaflops, making it the first supercomputer to surpass two exaflops of sustained computing power.

Design aimed at science and AI on one platform

Built at the National Supercomputing Center in Shenzhen, LineShine is designed to combine scientific computing and artificial intelligence on a single platform. Its developers created a software platform that integrates supercomputing and AI capabilities to allow researchers to make more efficient use of the hardware.

Core hardware and data movement

At the centre of the system are China’s self-developed LX2 processors, which support both traditional computing and AI workloads. The chips include what the developers describe as China’s first domestically developed high-bandwidth memory, which enables data to move about ten times faster than in conventional CPUs.

Network, storage and cooling

LineShine pairs raw compute with a proprietary high-speed interconnect network capable of linking up to two million ports and 100,000 nodes. Its storage system is designed to support both large-scale simulations and the massive datasets used in AI applications. The system uses a fully liquid-cooled design intended to improve energy efficiency while maintaining record-breaking performance.

Early scientific uses

Since entering operation, LineShine has supported research in atmospheric and ocean science, engineering simulation, materials research, drug discovery, brain science and scientific AI. The machine’s debut underscores the growing convergence of supercomputing and artificial intelligence in research infrastructure.

What this means for high-performance computing

LineShine’s achievement announced at an international conference for supercomputing marks China’s return to the forefront of high-performance computing and highlights a shift toward machines that handle both simulation workloads and data-intensive AI tasks on the same platform.