AMD to Collaborate With the U.S. Department of Energy to Build Two Cutting-Edge Supercomputers
According to a new Reuters report, AMD has secured a major partnership with the U.S. Department of Energy (DoE) to construct two new state-of-the-art supercomputers for academic and high-performance computing research. The collaboration represents a milestone for AMD, further solidifying its position in the supercomputing and AI markets while challenging NVIDIA’s dominance in the field.
The DoE projects, collectively valued at nearly $1 billion, will see AMD supplying its most advanced Instinct MI355X AI accelerators alongside a new MI430 variant that will power the next generation of large-scale computational systems. The partnership underscores the U.S. government’s intent to expand national computing capacity using AI-optimized, energy-efficient, and scalable architectures.
The first of the two systems, codenamed “Lux”, is scheduled to go online within the next six months, marking one of the fastest deployment timelines in recent DoE history. Lux will utilize AMD’s Instinct MI355X AI chips, the company’s newest accelerator designed for both AI model training and scientific workloads.
The Lux system will be developed in collaboration with Hewlett Packard Enterprise (HPE), Oracle, and Oak Ridge National Laboratory (ORNL), which previously hosted the Frontier supercomputer, currently one of the world’s fastest and also powered by AMD hardware. According to AMD CEO Lisa Su, the Lux deployment represents a record turnaround time for a national-scale supercomputing project.
A promotional image of the Frontier supercomputer at Oak Ridge features logos for the U.S. Department of Energy, Hewlett Packard Enterprise, and AMD, symbolizing the ongoing collaboration between the three.
The second system, named “Discovery”, is planned for delivery by 2028. It will feature a custom version of AMD’s Instinct MI430 AI accelerator, which is still in development and designed to push the boundaries of high-performance computing (HPC) and artificial intelligence workloads.
Project Discovery has reportedly been under review by the DoE since 2024, but with this announcement, AMD is now confirmed as the primary computing provider. The MI430 architecture is expected to build upon AMD’s next-generation chiplet design and advanced memory technologies, optimized for scientific research and national security applications.
The Department of Energy’s decision to extend its partnership with AMD reinforces the company’s growing reputation in the HPC space. The DoE has previously deployed AMD-based systems such as Frontier and El Capitan, both of which have demonstrated exceptional performance and scalability for AI-driven research.
While the Reuters report does not specify why AMD was chosen over NVIDIA, experts suggest that the DoE’s familiarity with AMD’s HPC architecture, combined with the company’s successful track record in energy-efficient performance and scalability, played a crucial role.
The partnership signals the DoE’s commitment to expanding its AI and supercomputing infrastructure through public-private collaborations. The department is reportedly seeking additional partnerships with other technology firms, which could include NVIDIA, to diversify and strengthen national computing resources.
For AMD, this partnership not only represents a technological achievement but also a significant strategic victory in its race against NVIDIA for leadership in the AI and HPC sectors.
Do you think AMD’s growing role in supercomputing could reshape the balance of power between major chipmakers in the AI era? Share your thoughts below.
