The United States has joined hands with chipmaker giant AMD in a $1 billion plan to build two powerful supercomputers that will support national research in areas such as nuclear energy, cancer treatment, and defense technology.
Energy Secretary Chris Wright told Reuters the supercomputers would speed up scientific progress in key areas by handling complex simulations and large volumes of data. He highlighted their importance in nuclear fusion research, describing the need to recreate the conditions at the center of the sun on Earth, which would allow for massive breakthroughs.
“We’ve made strong progress, but the instability of plasma means we need advanced systems to help reach fusion energy breakthroughs,” Wright said. He added that these machines will also be used to simulate molecular-level treatments for cancer and help manage the US nuclear weapons program.
First System to Launch in Six Months
The first supercomputer, named Lux, will be ready within six months. It will use AMD’s MI355X AI chips, along with AMD-made CPUs and networking components. The system is being developed with help from Hewlett Packard (HP) Enterprise, Oracle Cloud Infrastructure, and Oak Ridge National Laboratory (ORNL).
ORNL Director Stephen Streiffer said Lux will offer around three times more AI power than current systems. AMD CEO Lisa Su noted the speed of deployment was one of the fastest the company has seen for a computer of this size.
More Powerful Computer to Arrive in 2029
A second, more advanced system called Discovery will follow. It will feature AMD’s MI430 chips, designed to combine the functions of traditional supercomputing and AI processing. The machine is expected to be ready for use by 2029 and is also being developed by ORNL, AMD, and HPE.
A Department of Energy official said the agency will host both systems, while the companies will provide the hardware and share access to computing resources. The partnership marks what the official described as the beginning of a broader strategy to link DOE labs with private tech firms for future AI infrastructure projects.