The new technology will help manage the U.S. nuclear arsenal and accelerate the development of drugs to treat diseases such as cancer.
Published October 27, 2025
The United States has entered into a $1 billion partnership with Advanced Micro Devices (AMD) to build two supercomputers to tackle big scientific problems ranging from nuclear power to cancer treatment to national security.
Reuters first reported the new partnership, citing Energy Secretary Chris Wright and AMD CEO Lisa Su.
Recommended stories
list of 4 itemsend of list
The United States is building the two machines to ensure it has enough supercomputers at home to run increasingly complex experiments that require harnessing vast amounts of data processing power. This machine can accelerate the process of making scientific discoveries in areas where the United States is focused.
Energy Secretary Wright said the system would “accelerate” advances in nuclear and fusion energy, defense and national security technology, and drug development.
Scientists and companies are trying to recreate nuclear fusion, a reaction fueled by the sun, by disrupting lighter atoms in plasma gases under intense heat and pressure, releasing vast amounts of energy. “We’ve made great progress, but the plasma is unstable and we need to recreate the center of the sun on Earth,” Wright told Reuters.
“By using the calculations of these AI systems, we will make much faster progress. I believe we will have a practical path to harnessing fusion energy within the next two to three years.”
Wright said supercomputers could also help manage the U.S. nuclear arsenal and accelerate drug discovery by simulating cancer treatments at the molecular level.
“My hope is that within the next five to eight years, we will turn most cancers, many of which are now ultimately a death sentence, into manageable conditions,” Wright said.
“Speed and agility”
The plan is for the first computer, called Lux, to be built and operational within the next six months. It is based on AMD’s MI355X artificial intelligence chip, and the design will also include AMD’s central processing unit (CPU) and networking chips. The system was jointly developed by AMD, Hewlett Packard Enterprise (HP), Oracle Cloud Infrastructure, and Oak Ridge National Laboratory (ORNL).
AMD’s Su said the Lux deployment was the fastest deployment of a computer of this size he had seen.
“This is the speed and agility that we wanted in the U.S. AI effort,” Su said.
ORNL Director Steven Streifer said the Lux supercomputer will deliver approximately three times the artificial intelligence (AI) power of current supercomputers.
The second, more advanced computer, called Discovery, is based on AMD’s MI430 series AI chips, which are tuned for high-performance computing. This system will be designed by ORNL, HP, and AMD. Discovery is expected to be delivered in 2028 and operationally ready in 2029.
Streifer said he expected a big improvement, but couldn’t predict just how much computational power would be gained.
MI430 is a special version of AMD’s MI400 series, which combines key features of traditional supercomputing chips with the ability to run AI applications, Su said.
The U.S. Department of Energy (DOE) will host the computers, companies will provide the machinery and capital investment, and both sides will share computing power, DOE officials said.
The two supercomputers, based on AMD chips, are intended to be the first of their kind in partnership with private industry and DOE laboratories across the country, officials said.
AMD stock was up 0.7% as of 12:15 p.m. New York (16:15 p.m. Japan time).
