The U.S. government has entered into a $1 billion partnership with Advanced Micro Devices, Inc. (NASDAQ:AMD) to develop two next-generation supercomputers, a move aimed at cementing America’s technological leadership in high-performance computing.
The initiative underscores Washington’s growing commitment to expanding computational capacity for scientific discovery and national security.
The two machines, Energy Secretary Chris Wright and AMD CEO Lisa Su told Reuters, will have the capability to address some of the world’s most complex challenges, from advancing nuclear energy research and developing new cancer treatments to enhancing the nation’s security infrastructure.
Also Read: Amazon Takes On Nvidia With Cheaper AI Supercomputers, Servers
Next-Generation Systems
The first system, named Lux, is slated to come online within the next six months. It will integrate AMD’s central processors (CPUs) and networking technology and is built around the company’s MI355X artificial intelligence chips.
Lux is being co-developed by AMD, Hewlett Packard Enterprise Co. (NYSE:HPE), Oracle Corp. (NYSE:ORCL) Oracle Cloud Infrastructure, and the Oak Ridge National Laboratory (ORNL).
Su called the Lux system the fastest large-scale deployment she has seen in her career. ORNL Director Stephen Streiffer noted that Lux will deliver approximately three times the AI capacity of current supercomputers.
AMD’s MI430 AI chips will power the second, more advanced system, Discovery. Developed by ORNL, HPE, and AMD, Discovery is expected to be completed in 2028 and operational by 2029.
This investment underscores a broader industry push, as chipmakers aggressively develop supercomputers to meet the surging demand for AI and HPC (High-Performance Computing), simultaneously boosting economic growth and fortifying their supply chains.
Competitive Landscape
AMD rival Nvidia Corp. (NASDAQ:NVDA) recently stepped up its efforts, launching its compact DGX Spark personal AI supercomputer in October.
Nvidia’s move aims to deliver data center-level computing power to desktops globally for researchers, developers, and students.
Starting Wednesday, October 15, Nvidia began accepting online orders for the DGX Spark through its website and select U.S. retailers.
Powered by the GB10 Grace Blackwell Superchip, Spark can run inference on AI models with up to 200 billion parameters and fine-tune models up to 70 billion parameters, capabilities previously reserved for massive cloud systems.
The launch coincides with an AI computing boom, which has seen Nvidia’s market value surge past $4.5 trillion, with CEO Jensen Huang forecasting $500 billion in annual AI data center spending by 2030.
Price Action: AMD stock was trading higher by 1.28% to $256.17 at last check Monday.
Read Next:
Image via Shutterstock
© 2025 Benzinga.com. Benzinga does not provide investment advice. All rights reserved.
