AMD and Department of Energy Ink $1 Billion Deal for AI Supercomputers

AMD has secured a significant $1 billion agreement with the U.S. Department of Energy (DOE) to spearhead the development of two advanced supercomputers, codenamed Lux and Discovery. This ambitious project, undertaken in collaboration with industry giants Oracle and Hewlett Packard Enterprise (HPE), will see both powerful machines housed at the renowned Oak Ridge National Laboratory (ORNL) in Oak Ridge, Tennessee. Lux is scheduled to become operational in early 2026, with Discovery expected to follow its launch in 2029.

Building on a Legacy of High-Performance Computing

These forthcoming supercomputers represent an evolution of the high-performance computing initiatives previously undertaken. They build directly on the foundational work that went into the Frontier supercomputer, which also resides at ORNL and held the title of the world's fastest supercomputer until the debut of El Capitan at Lawrence Livermore National Laboratory last year. AMD's involvement in developing both Frontier and El Capitan underscores its established expertise and ongoing partnership with the U.S. government on such critical technological ventures.

Lux: The Nation's First AI Factory for Science

The first of the two, Lux, is being heralded as the nation's inaugural "AI Factory" dedicated to advancing science, energy, and national security. It is specifically engineered to train, fine-tune, and deploy sophisticated AI foundation models, with the ultimate goal of accelerating scientific discovery and fostering engineering innovation. Its cutting-edge architecture is meticulously optimized for data-intensive and model-centric workloads, making it a pivotal asset for AI-driven research.

Discovery: Enhanced Performance and Energy Efficiency

Following Lux, the Discovery supercomputer is set to introduce a "Bandwidth Everywhere" design. This innovative approach promises substantial improvements over the Frontier supercomputer in terms of both performance and energy efficiency, aiming to deliver greater computing output at a comparable operational cost. Discovery's immense processing capabilities are destined to drive breakthroughs across a wide spectrum of scientific disciplines. Its applications will span critical areas such as energy, biology, advanced materials, national security, and manufacturing innovation, playing a crucial role in the design of next-generation reactors, batteries, catalysts, semiconductors, and essential materials.