AMD has secured a partnership with the U.S. Department of Energy to build two next-generation supercomputers, Lux and Discovery, targeting scientific and national security challenges ranging from nuclear fusion research to cancer treatments. Energy Secretary Chris Wright said the effort will supercharge progress in nuclear energy, defense technologies and drug discovery, accelerating breakthroughs that typically take decades. The Lux system is slated to go online within six months, reflecting what AMD CEO Lisa Su described as the fastest deployment of this scale the company has ever executed.
Lux is being co-developed by AMD, Hewlett Packard Enterprise, Oracle Cloud Infrastructure and Oak Ridge National Laboratory. It will be powered by AMD’s Instinct MI355X Artificial Intelligence accelerators alongside AMD CPUs and networking chips. Oak Ridge National Laboratory Director Stephen Streiffer said Lux will deliver roughly three times the Artificial Intelligence performance of today’s top supercomputers, positioning it as a major step forward for workloads that demand rapid model training and complex simulations.
The second system, Discovery, will follow later in the decade and is built on AMD’s upcoming MI430 Artificial Intelligence chips. The project is expected to be completed by 2028 and become operational in 2029. Both machines are part of a broader U.S. push to maintain leadership in high-performance computing and Artificial Intelligence by expanding capacity for scientific computing and mission-critical research.
According to Reuters, the Department of Energy’s choice reflects confidence in AMD’s high performance computing record, including Frontier and El Capitan, the world’s fastest supercomputers that also run on AMD hardware. Wright said Lux and Discovery will enable advanced simulations of plasma dynamics for fusion energy and molecular interactions for drug development, bringing massively faster progress toward harnessing fusion and treating severe diseases. Together, the systems mark a milestone for AMD and U.S. scientific computing and build momentum for AMD’s upcoming MI400 accelerators series.
