Top view of an AMD Ryzen 5 7600 processor, part of the company’s next-generation desktop lineup. AMD’s chip technology continues to expand from consumer computing into advanced AI and supercomputing applications.

In a new $1 billion collaboration, Advanced Micro Devices will work with the U.S. Department of Energy to create two powerful supercomputers aimed at solving some of science’s most complex challenges. The effort, unveiled by Energy Secretary Chris Wright alongside AMD CEO Lisa Su, will expand the nation’s capacity for high-performance computing and artificial intelligence research.

The project’s goal is to provide scientists with tools capable of modeling vast systems—from nuclear fusion reactions to molecular-level drug design—with unprecedented speed and accuracy. Secretary Wright said the upcoming machines will enhance national security programs, advance clean-energy development, and speed up the discovery of new medical treatments.

“These computers give us the ability to test ideas that used to take years in the lab,” Wright said. “They’ll help us refine fusion experiments, simulate plasma behavior, and explore how cancer cells respond to therapies before those treatments ever reach a patient.”

The first of the two systems, called Lux, is set to begin operation within six months. It will incorporate AMD’s latest artificial intelligence processors combined with specialized networking and data-handling components. The project is being built in partnership with Hewlett Packard Enterprise, Oracle’s cloud division, and Oak Ridge National Laboratory in Tennessee. Researchers expect Lux to triple the A.I. performance currently available at U.S. laboratories.

Su said the rapid development timeline represents a new model for collaboration between government and industry. “This is the kind of pace that’s needed to stay ahead in artificial intelligence and supercomputing,” she said.

The second system, named Discovery, will arrive later this decade and will rely on AMD’s forthcoming MI430 series processors, designed for hybrid workloads that merge scientific simulation with deep learning. Discovery’s installation and testing are scheduled for completion by 2029, again under Oak Ridge’s direction.

Under the terms of the collaboration, the Energy Department will host both supercomputers while AMD and its partners supply the hardware and investment. Computing resources will be shared across government research, academic institutions, and industry partners.

A department representative said this approach is part of a long-term plan to pair U.S. technology firms with federal research centers, accelerating innovation in fields ranging from energy and materials science to national defense. By combining public research infrastructure with private engineering expertise, officials hope to open a new era of scientific discovery powered by artificial intelligence and advanced computing.

Image is licensed under the GNU Free Documentation License and was created by Smial (talk).