AMD has said it supplies components to four of the world’s 10 fastest supercomputers, as the company outlined its expanding role in high-performance computing and AI infrastructure.
The chipmaker appears in 177 systems on the Top500 list, representing 35% of ranked machines, and in 26 of the 50 most energy-efficient systems on the Green500. Its processors also feature in El Capitan and Frontier, the top two US Department of Energy systems.
Cloud providers are continuing to broaden their hardware choices. Microsoft has made its Azure HBv5 virtual machines, based on 5th-generation Epyc CPUs, generally available, and Google’s H4D VMs — using the same generation — have shown higher performance than their predecessors on a range of HPC workloads.
AMD hardware will also underpin two new US government systems, Lux and Discovery, which will be deployed at Oak Ridge National Laboratory. Lux is expected to enter service in early 2026. Both systems will mix Epyc CPUs, Instinct accelerators, and Pensando networking to support research in areas such as energy, materials, and national security.
Alongside its hardware, AMD has introduced the Enterprise AI Suite, an open-source platform designed to help organisations manage AI workloads in Kubernetes-based environments.
Researchers at Lawrence Livermore National Laboratory recently used El Capitan to run what they described as the largest and fastest protein structure prediction workflow to date, carried out with AMD and Columbia University.
AMD said these projects reflect its continued presence in supercomputing and AI, though several of the systems it highlighted will not be fully operational for some time.
