AMD powers four of the world’s 10 fastest supercomputers and dominates the Green500 efficiency rankings, while previewing its next-generation MI430X GPU to deliver over 200 TFLOPs of scientific computing performance.
In a press release dated 13 June 2026, AMD announced that its technology now powers four of the world’s 10 fastest supercomputers. AMD powers 191 systems across the TOP500 list, reflecting an 11% increase year-over-year and securing 41% of all new entrants.
AMD Instinct GPUs account for 42% of the total floating-point operations per second (FLOPs) delivered by the top 10 systems combined. The specific AMD-powered supercomputers highlighted in the top 10 include El Capitan (No. 2), Frontier (No. 3), and HPC7 (No. 6).
According to the Green500 rankings, AMD powers 56% of the top 50 systems, demonstrating a massive footprint in performance-per-watt efficiency. It claims four of the world’s 10 most energy-efficient supercomputers: Otus (No. 5), Capella (No. 6), AMD Ouranos (No. 9), and Portage (No. 10).
The first systems running AMD Instinct MI355X GPUs debuted at the University of Cambridge (ranking 67 and 68), forming part of Zenith AI, the UK’s largest AI-for-science platform. Meanwhile, the LUMI supercomputer in Finland sits at No. 11, remaining a vital asset for European AI research.
Looking ahead, France’s national computing agency, GENCI, is developing Alice Recoque—the country’s first exascale supercomputer. It will function as an “AI Factory” powered by 6th Gen AMD EPYC CPUs and next-generation AMD Instinct MI430X GPUs. AMD previewed this upcoming MI430X GPU, engineered for heavy double-precision (FP64) scientific computing like climate modelling. It is projected to deliver over 200 teraflops (TFLOPs) of native FP64 performance, setting a new industry benchmark.
