AMD to build two more supercomputers at Oak Ridge National Labs

News
Oct 30, 20252 mins

The two AMD-based supercomputers will be geared towards discoveries and science, energy, and national security.

flash supercomputers 16_9
Credit: Timofeev Vladimir - shutterstock.com

The collaboration between AMD and the U.S. Department of Energy (DOE) Facility at the Oak Ridge National Laboratory (ORNL) continues, with two more supercomputers planned to join the two that already are deployed.

The two existing systems at ORNL are Frontier, an all-AMD design, and Summit, an IBM/NVIDIA design. They will be joined in the next few years by Lux AI and Discovery. Discovery and Lux will be DOE flagship supercomputers designed to drive breakthroughs in science, energy and national security. The two machines represent a combined $1 billion investment of private and public funding. 

 [ RelatedMore AMD news and insights ]

The Lux AI supercomputer is being jointly developed by ORNL, AMD, Oracle Cloud Infrastructure (OCI), and HPE. It will be powered by AMD EPYC CPUs, Instinct MI355X GPUs, and Pensando networking technologies. Deployment is scheduled for early 2026, positioning it as the first U.S. AI Factory supercomputer supporting both public and private sector collaborations.

Lux is engineered to train, refine, and deploy AI foundation models that accelerate scientific and engineering progress. Its advanced architecture supports data-intensive and model-centric workloads, thereby enhancing AI-driven research capabilities.

Discovery differs from Lux in that it uses Instinct MI430X GPUs instead of the 300 series. The MI400 Series is designed for sovereign AI and scientific applications. It allows the U.S. to develop and deploy AI models on domestically manufactured systems to safeguard national data and competitive standing in science.

Discovery is purpose-built for AI and scientific applications, with peak bandwidth to maximize memory, node, and network bandwidth to enhance performance in science and AI workloads. Discovery will use HPE’s newly launched HPE Cray Supercomputing GX5000 platform, which is designed for converged AI/HPC workloads, while Frontier uses the older HPE Cray EX4000. The two programming environments are compatible and transferrable however, so anything developed on the Frontier is transferable to the Discovery systemchines.

Andy Patrizio is a freelance journalist based in southern California who has covered the computer industry for 20 years and has built every x86 PC he’s ever owned, laptops not included.

Andy writes the Data Center Explorer blog for Network World. His work has appeared in a variety of publications, including Tom's Guide, Wired, Dr. Dobbs Journal, Tech Target, Business Insider, and Data Center Knowledge. Earlier in his career, he held editorial positions at IT publications like InternetNews, PC Week and InformationWeek.

Andy holds a BA in Journalism from the University of Rhode Island.

More from this author