China’s LineShine Supercomputer Claims Global Top Rank, Dethroning U.S.-Built Systems After 10 Years

China has officially overtaken the United States to field the world’s fastest operational supercomputer, marking its return to the top of global supercomputing leaderboards after nearly a decade outside first place.

Housed at the National Supercomputing Center in Shenzhen, the newly debuted system named LineShine outperformed the U.S. El Capitan platform — the long-standing prior leader of the TOP500 ranking — across all measured computing power metrics. Prior to being displaced, El Capitan, installed at the Lawrence Livermore National Laboratory in California, had held the number one spot on the bi-annual TOP500 list since 2024. Independent verified benchmark results confirm LineShine’s total processing capacity exceeds its U.S. predecessor by more than 20%.

---

Breakdown of LineShine’s Unconventional, Fully Indigenous Design

First established in 1993, the TOP500 ranking is the global gold standard for supercomputing performance, updated every six months via a set of standardized, peer-vetted tests that evaluate far more than theoretical peak speed: it also measures real-world task performance, energy efficiency, and long-term operational stability. For most of its 30+ year history, the ranking has been dominated by systems developed and assembled in the U.S.

LineShine’s record-breaking performance far exceeds the standard exascale performance threshold: it delivers 2,198 exaflops of computing throughput, capable of running more than 2 quintillion individual operations per second, while drawing roughly 42.2 megawatts of power.

What makes the platform particularly unprecedented among top-tier next-generation supercomputers is its total avoidance of graphics processing units (GPUs), the specialized AI-accelerator chips that power nearly all of the world’s highest-performing modern supercomputers. Instead, LineShine runs entirely on standard central processing units (CPUs) — the same general-purpose chips found in consumer smartphones, desktops, and laptops, but rarely deployed at massive scale for large scientific computing workloads.

Even more notably, every layer of LineShine’s full tech stack is 100% developed and manufactured by Chinese domestic tech firms:

• Its core architecture is built on the homegrown LingKun platform, with a total of ~45,000 custom LX2 processors, each packing 304 cores running at a 1.55GHz clock speed

• Nodes across the massive system are connected via the LingQi high-speed interconnect network, purpose-built to cut down data transfer latency and speed up cross-node workload syncing

• The entire supercomputer runs on Kylin OS, a Linux-based domestic operating system already widely deployed across China’s government and scientific computing infrastructure.

---

The Milestone As A Clear Signal Amid Ongoing U.S.-China Tech Competition

LineShine’s top ranking is far more than a symbolic technical win, widely interpreted as a public demonstration that China’s domestic tech industry can produce world-leading performance even when locked out of access to core U.S. advanced technologies.

Years of escalating U.S. export controls, first introduced during Donald Trump’s first presidential term and extended throughout Joe Biden’s administration, were explicitly designed to slow China’s high-tech progress by blocking shipments of advanced computing components, software, and development platforms. Restrictions have grown even stricter under Trump’s current second administration, via expanded steep tariffs and broad new bans on imports of AI-focused GPUs, cutting-edge semiconductors, and other core components that form the backbone of most top-tier global supercomputers. Beijing has rolled out reciprocal targeted trade restrictions in response.

These sweeping barriers pushed Chinese engineering and research teams to pour resources into developing entirely new, indigenous supercomputing architectures that could deliver performance on par with the world’s top U.S. systems, without relying on restricted imported parts. LineShine’s historic top spot is the first major public payoff of that multi-year push.

This report was originally published in Spanish on WIRED en Español, and adapted for English audiences.