The Significance of AMD’s Collaboration with the DOE for Enterprise AI Success

The U.S. Department of Energy (DOE) and AMD have joined forces to collaborate on two cutting-edge AI supercomputers at Oak Ridge National Laboratory (ORNL). This partnership is part of a broader AI strategy aimed at advancing research in science, energy, and national security, while also bolstering the nation’s standing in high-performance computing.

With a combined investment of approximately $1 billion from both public and private sources, these two supercomputers will be integrated into a secure national computing network focused on supporting AI research using US-built infrastructure that adheres to standards. This project demonstrates how a cohesive AI strategy can align with national objectives in innovation, energy efficiency, and data governance.

Dr. Lisa Su, the Chair and CEO of AMD, expressed pride and honor in collaborating with the DOE and ORNL to accelerate America’s foundation for science and innovation. She highlighted that the systems will leverage AMD’s high-performance and AI computing technologies to advance critical research priorities in the U.S. related to science, energy, and medicine.

Lux AI, set to launch in early 2026, will be the nation’s first “AI Factory” dedicated to training and deploying advanced AI models for applications in science, energy, and security. Developed in partnership with ORNL, AMD, Oracle Cloud Infrastructure, and Hewlett Packard Enterprise, Lux will utilize AMD Instinct MI355X GPUs, EPYC CPUs, and Pensando networking to tackle data-intensive AI tasks. Its design aims to accelerate research efforts in areas such as energy systems, materials, and medicine, while facilitating collaboration among multiple groups in a secure environment.

The Discovery system, scheduled for deployment in 2028, will become the DOE’s next flagship supercomputer at Oak Ridge. Powered by AMD’s upcoming “Venice” EPYC processors and MI430X GPUs, specifically designed for AI and scientific computing, Discovery incorporates a “Bandwidth Everywhere” architecture to enhance memory and network performance without increasing power consumption. This design enables the system to process larger volumes of data and run complex models efficiently, addressing the challenge of maintaining energy costs faced by many large data centers today. Drawing lessons from the Frontier supercomputer, the world’s first exascale computer, Discovery ensures seamless migration of existing applications to the new platform.

U.S. Energy Secretary Chris Wright emphasized the importance of innovative partnerships to win the AI race, bringing together top minds and industries in American technology and science. The new systems represent a practical approach to computing partnerships that enhance the country’s competitiveness through collaborative innovation.

ORNL Director Stephen Streiffer underlined how Discovery will drive scientific innovation to new heights, combining high-performance computing and AI to expedite the translation of research problems into real-world solutions. The involvement of companies like AMD, HPE, and Oracle in building and supporting these systems underscores the significance of partnerships in driving AI innovation and long-term strategy.

Antonio Neri, President and CEO of HPE, highlighted that the collaboration will enable Oak Ridge to achieve unprecedented productivity and scale, while Oracle’s Executive Vice President, Mahesh Thiagarajan, emphasized the delivery of high-performance AI infrastructure to support the Lux AI cluster’s co-development. Lux and Discovery will empower the DOE to run large-scale AI models, advancing understanding in energy, biology, materials science, and national defense. Discovery will also contribute to the design of next-generation technologies such as batteries, reactors, semiconductors, and critical materials.

For organizations, these supercomputing systems underscore the potential of AI strategy and high-performance computing to drive faster research, enhance efficiency, and ensure secure data management. They demonstrate that performance gains can be achieved without escalating energy consumption. The DOE’s collaborative approach with technology providers presents a model that private enterprises can emulate, combining expertise across sectors to develop shared infrastructure while maintaining data control.

As AI workloads continue to grow, both public and private organizations will need to design systems that strike a balance between power, performance, and governance. The Lux and Discovery projects exemplify this balance in action – fostering an open, collaborative environment tailored to support large-scale discoveries. These initiatives serve as a testament to how a forward-thinking AI strategy can transform infrastructure into a sustainable competitive advantage.