www.magazine-industry-usa.com
28
'25
Written on Modified on
AMD to Power U.S. Sovereign AI Supercomputers at Oak Ridge National Lab
Lux and Discovery systems will advance America’s open, secure AI and HPC leadership.
www.amd.com

AMD and the U.S. Department of Energy (DOE) announced two next-generation systems at Oak Ridge National Laboratory (ORNL) designed to expand America’s leadership in artificial intelligence (AI) and high-performance computing (HPC), the Lux AI supercomputer and the Discovery supercomputer. Discovery and Lux will be DOE flagship supercomputers designed to drive breakthroughs in science, energy, and national security. Both systems directly support the U.S. AI Action Plan by accelerating AI-enabled science, strengthening national competitiveness, and advancing secure, sovereign AI infrastructure for the nation.
Together with the U.S. government, ORNL and industry partners, AMD will deliver advanced computing platforms that help national researchers, agencies, and innovators tackle pressing challenges in energy, medicine, health, and national security. When fully deployed, the Lux and Discovery systems will represent a combined $1billion investment of private and public funding which will enable the DOE to build a secure, federated and standards-based infrastructure for sovereign U.S. AI and science.
The Lux AI supercomputer – co-developed by ORNL, AMD, Oracle Cloud Infrastructure (OCI) and HPE and powered by AMD Instinct™ MI355X GPUs, AMD EPYC™ CPUs and AMD Pensando™ advanced networking technologies – will be deployed in early 2026, making it the first US AI Factory supercomputer setting the stage for novel public and private partnerships. The Lux system will address the immediate needs to expand the DOE’s AI leadership and accelerate progress and innovation in the areas of AI, energy research, materials, medicine and advanced manufacturing.
The Discovery supercomputer deepens the collaboration between the DOE, ORNL, HPE and AMD. At the heart of Discovery are next-gen AMD EPYC CPUs, codenamed “Venice,” and AMD Instinct™ MI430X GPUs— a new MI400 Series accelerator engineered specifically for sovereign AI and scientific computing. MI430X extends the leadership AMD architecture for AI and HPC, enabling the U.S. to train, simulate, and deploy AI models on domestically built systems that protect national data and scientific competitiveness.
Lux AI Supercomputer: Advancing U.S. AI Leadership
Lux at ORNL is the nation’s first dedicated AI Factory for science, energy, and national security—purpose-built to train, fine-tune, and deploy AI foundation models that will accelerate discovery and engineering innovation. Lux is designed to accelerate AI-driven science through its advanced architecture, optimized for data-intensive and model-centric workloads.
As a critical resource for the DOE and the National Laboratory system, Lux will give the United States an early and decisive advantage in deploying frontier AI capabilities that transform the pace and scale of scientific innovation. By dramatically expanding the nation’s capacity for scientific AI, Lux strengthens U.S. leadership, accelerates AI-for-Science, and helps ensure that the breakthroughs of the AI era are discovered, developed, and deployed in America.
Lux will unleash a new generation of scientific capabilities across the DOE’s mission to, reveal nature’s hidden structure, discover breakthrough materials, accelerate energy innovation, transform biology and biosecurity, and advance national security and resilience.
Discovery: America’s Next Leap in Scientific Computing
Purpose‑built for AI and science, Discovery features a “Bandwidth Everywhere” design that allows science and AI applications to run efficiently and deliver productive results. Importantly, Frontier laid the foundation for this work, and together with HPE Cray Supercomputing GX5000 – its next-generation supercomputing platform – Discovery builds on that success with a consistent programming environment to ensure an easy transition for users.
Discovery will be the DOE’s next flagship supercomputer at ORNL, arriving in 2028, and will extend U.S. leadership in HPC and AI. Advancing America’s innovation agenda, Discovery reflects a strong, public‑private commitment to keep the United States at the forefront of AI and scientific innovation.
Building on the breakthroughs of Frontier—the world’s first exascale system—Discovery is designed to deliver dramatically higher performance, energy efficiency, and AI capabilities. Discovery will drive breakthroughs in energy, biology, advanced materials, national security, and manufacturing innovation. It will help design next-generation reactors, batteries, catalysts, semiconductors, and critical materials.
Key Points of Discovery:
- Bandwidth Everywhere: Impressive memory capacity, node, and global network bandwidth compared to first-generation exascale machines, accelerating science and AI applications.
- Seamless User Transition: Applications built for Frontier transition seamlessly as Discovery continues the successful programming environment.
- More Compute, Similar Energy: More AI and scientific output at comparable power costs.
- Open Standards: Discovery is built on open-source software, open standards, and open science to provide the foundation for AI sovereignty.
ORNL is expected to take delivery of Discovery in 2028, with user operations expected in 2029. Once online, Discovery will serve as a cornerstone of the American AI Stack, advancing the science, security, and innovation goals that define the nation’s next decade.

