U.S. partners with AMD in $1 billion project to build next-generation supercomputers

The U.S. Department of Energy’s partnership with AMD, HPE, Oracle Cloud, and ORNL aims to build powerful supercomputers that will empower American scientists to address the decade’s most complex scientific and technological challenges

In a major leap toward advancing national scientific capabilities, the United States has partnered with Advanced Micro Devices (AMD) in a $1 billion initiative to build two state-of-the-art supercomputers. The systems will drive breakthroughs in nuclear energy, cancer research, artificial intelligence (AI), and national security, according to U.S. Energy Secretary Chris Wright and AMD CEO Lisa Su.

The collaboration, announced through the U.S. Department of Energy (DOE), brings together AMD, Hewlett Packard Enterprise (HPE), Oracle Cloud Infrastructure, and Oak Ridge National Laboratory (ORNL). These new high-performance systems will equip American scientists with the computational power needed to tackle some of the most complex scientific challenges of the decade.

Accelerating scientific innovation through AI computing

Energy Secretary Wright said the two supercomputers would “supercharge” advancements in areas such as nuclear power, fusion energy, and medical research. The powerful machines will enable researchers to simulate extreme conditions like those at the center of the sun — critical to developing fusion-based clean energy. “We’ve made great progress, but plasmas are unstable, and we need to recreate the center of the sun on Earth,” Wright said. “With AI-driven computation, we can make faster progress toward practical fusion energy within the next few years.”

The supercomputers will also accelerate drug discovery by allowing scientists to simulate and analyze cancer treatments at the molecular level. Wright added that with such computational advances, several fatal cancers could become manageable within the next five to eight years. Additionally, the systems will help the U.S. maintain and secure its nuclear arsenal more effectively.

Introducing Lux and Discovery supercomputers

The first system, Lux, is scheduled to be operational within six months. It will feature AMD’s MI355X AI accelerators, CPUs, and networking chips, delivering nearly three times the AI capacity of current leading supercomputers. ORNL Director Stephen Streiffer said Lux represents the fastest deployment of its scale in U.S. computing history.

The second system, Discovery, will follow by 2029, powered by AMD’s next-generation MI430 AI chips optimized for high-performance computing and AI workloads. Both machines will be hosted by the DOE, symbolizing a new model of public-private collaboration aimed at strengthening U.S. leadership in AI-driven scientific research and innovation.