AMD To Power U.S. Ai Factory Supercomputers At Oak Ridge

AMD has partnered with the US Department of Energy (DOE) and Oak Ridge National Laboratory (ORNL) to deliver two new supercomputers designed to advance sovereign Ai and high-performance computing.

The systems, named Lux and Discovery, will be deployed at ORNL as part of a broader initiative to strengthen U.S. leadership in science, energy and national security. The combined investment from public and private funding is expected to total around $1 billion.

Lux: First Ai Factory For Science

Lux is set to become the country’s first dedicated Ai Factory for science when deployed in 2026. Co-developed by ORNL, AMD, Oracle Cloud Infrastructure and HPE, the system will run on AMD Instinct MI355X GPUs, AMD Epyc CPUs and AMD Pensando networking technologies.

The project is designed to train and deploy Ai foundation models that can accelerate research in energy, medicine, materials and advanced manufacturing.

“Lux will give the United States an early advantage in deploying frontier Ai capabilities that transform the pace and scale of scientific innovation,” ORNL said in its launch announcement.

Mahesh Thiagarajan, Executive Vice President at Oracle Cloud Infrastructure, said Oracle’s role in the partnership will help deliver sovereign, high-performance Ai infrastructure.

“Oracle is honored to work alongside the Department of Energy to help drive breakthroughs in science, energy and national security,” Thiagarajan said.

Discovery: Next Flagship Supercomputer

The second system, Discovery, will become ORNL’s flagship supercomputer in 2028. At its core will be AMD’s next-generation Epyc CPUs, codenamed “Venice,” and AMD Instinct MI430X GPUs, which are part of the new MI400 Series accelerators.

The DOE expects Discovery to extend U.S. capabilities in Ai and scientific computing, with applications across biology, materials science, energy innovation and national security.

Chris Wright, U.S. Secretary of Energy, said Discovery represents a shift in how large-scale computing partnerships are being managed.

“Winning the Ai race requires new and creative partnerships that bring together the brightest minds and industries American technology and science has to offer,” Wright said. “Working with AMD and HPE, we’re bringing new capacity online faster, proving that America leads when private-public partners build together.”

Supporting National Ai Strategy

Both systems are part of the U.S. Ai Action Plan, which seeks to build a secure, standards-based infrastructure for sovereign Ai and science.

Dr. Lisa Su, Chair and CEO of AMD, said the projects reflect the role of public-private partnerships in advancing national research.

“Discovery and Lux will leverage AMD’s high-performance and Ai computing technologies to advance critical U.S. research priorities in science, energy and medicine,” Su said.

Antonio Neri, President and CEO of HPE, described Discovery as a step forward in converging Ai and high-performance computing.

“HPE’s newest supercomputing solutions harness converged Ai and HPC architectures, enabling the lab to achieve unprecedented productivity and scale,” Neri said.

Building On Frontier

Discovery will build on the experience gained from Frontier, ORNL’s current exascale supercomputer, and offer a programming environment that makes migration easier for researchers.

The system will feature higher bandwidth, greater memory capacity and more efficient power use compared to earlier exascale machines. ORNL expects to take delivery in 2028, with user operations beginning in 2029.

ORNL Director, Stephen Streiffer, said the integration of Ai with high-performance computing would have a direct impact on how quickly researchers can achieve results.

“With Discovery, the integration of high-performance computing and Ai promises breakthroughs at the speed and scale necessary for continued U.S. leadership in an increasingly competitive global environment,” Streiffer said.

Long-Term Goals

Once operational, Discovery will serve as a cornerstone of what the DOE is calling the American Ai Stack. This national infrastructure will support both scientific research and applications with direct relevance to national security and resilience.

The DOE said the systems will help design next-generation reactors, semiconductors, batteries and other critical technologies.

By combining Ai and HPC, the Lux and Discovery supercomputers are intended to establish a secure, sovereign computing environment to underpin the country’s science and innovation agenda for the next decade.

Last Updated on October 28, 2025 by Nick Ross