According to Reuters, U.S. Energy Secretary Chris Wright and AMD CEO Lisa Su announced that the U.S. Department of Energy and AMD have reached a $1 billion partnership agreement to build two supercomputers to tackle major scientific challenges, ranging from nuclear energy and cancer treatment to national security.

The United States is working to build these two supercomputers to ensure the nation has sufficient computing power to run increasingly complex scientific experiments that require processing vast amounts of data and performing intensive computational analysis. The new systems will accelerate scientific discovery in key U.S. science and technology areas.

Wright said these supercomputers will "significantly advance" the development of nuclear and fusion energy, defense and national security technologies, and new drug research and development. Scientists and companies are currently trying to simulate the nuclear fusion reactions inside the sun by compressing light atoms in plasma gas at extremely high temperatures and pressures, thereby releasing enormous amounts of energy.

"We've made significant progress, but the plasma is extremely unstable, so we have to recreate the conditions at the core of the Sun here on Earth," Wright said. "I believe that with the computing power of these AI systems, we'll have a practical path to practical fusion energy within the next two to three years, and scientific progress will accelerate dramatically."

Wright also noted that these supercomputers will help manage the U.S. nuclear arsenal and accelerate drug development by simulating cancer treatments at the molecular level.

"I hope that within the next five to eight years, most cancers, which are currently considered terminal, will be manageable chronic diseases," Wright said.

The first supercomputer, named "Lux," is scheduled to be up and running within the next six months. The system will be based on AMD's latest MI355X AI chip, along with AMD's own central processing unit (CPU) and networking chips. The project is a joint venture between AMD, Hewlett Packard Enterprise (HPE), Oracle Cloud Infrastructure, and Oak Ridge National Laboratory (ORNL).

AMD CEO Lisa Su said Lux ​​is the fastest supercomputer deployment of its scale she has ever seen. "This is exactly the speed and agility we're striving for to support the advancement of AI in the United States," Su said.

ORNL Director Stephen Streiffer explained that the Lux supercomputer is expected to deliver approximately three times the AI ​​computing power of existing systems of the same type.

The second, more advanced supercomputer, named Discovery, will be powered by AMD's MI430 series AI chip, optimized for high-performance computing. Jointly designed by ORNL, HPE, and AMD, the system is expected to be delivered in 2028 and operational in 2029.

Streiffer stated that while a significant performance leap is expected, the exact magnitude of the increase in computing power cannot be accurately predicted at this time.

Su explained that the MI430 is a specialized variant of the MI400 series, combining key features of traditional supercomputing chips with the functionality required to run AI applications, offering both general-purpose computing and AI acceleration capabilities.

A U.S. Department of Energy official stated that the two supercomputers will be hosted by the Department of Energy, while the partner companies will provide equipment and capital investment, and both parties will share computing resources. The official also revealed that these two supercomputers, built on AMD chips, will be the beginning of a series of public-private partnerships across the United States, and that such collaborations will be promoted in the future between more private companies and the Department of Energy's national laboratories.