AMD Secures $1 Billion Energy Department Deal to Build Advanced Supercomputers

TLDR

• AMD won a $1 billion contract with the U.S. Department of Energy to construct two supercomputers for scientific research
• Lux supercomputer launches in six months using MI355X chips, Discovery arrives in 2029 with MI430 processors
• Systems will support fusion energy research, cancer drug development, and national security applications
• IBM separately tested quantum algorithms on AMD chips with successful results running ten times faster than required
• AMD shares climbed 2.67% to $259.67 following the announcement

Advanced Micro Devices has secured a $1 billion partnership with the U.S. Department of Energy to develop two supercomputers. The systems will advance research in clean energy, medical treatments, and defense technologies.

Energy Secretary Chris Wright stated the machines will enable scientists to tackle computational problems that exceed current capabilities. The partnership represents one of AMD’s largest government contracts. The announcement came just days before AMD’s third-quarter earnings report on November 4.

The first system, Lux, will begin operations within six months. It uses AMD’s MI355X artificial intelligence chips combined with the company’s central processors and networking technology. AMD is collaborating with Hewlett Packard Enterprise, Oracle Cloud Infrastructure, and Oak Ridge National Laboratory on the project.

Advanced Micro Devices, Inc., AMD

Oak Ridge National Laboratory Director Stephen Streiffer confirmed Lux will provide three times the AI capacity of existing supercomputers. AMD CEO Lisa Su described the timeline as the fastest deployment she has witnessed for a computer of this scale.

Applications in Energy and Medicine

Wright explained the supercomputers will help scientists study plasma stability for fusion energy development. Fusion research aims to replicate the sun’s energy production process on Earth. The energy secretary projects the computing power will create viable fusion energy pathways within two to three years.

The systems will also model cancer treatments at the molecular level to accelerate drug discovery. Wright expects many cancers could become manageable conditions within five to eight years. The computers will support management of the U.S. nuclear weapons arsenal as well.

Second System and Quantum Computing Development

The Discovery supercomputer will launch in 2029 using AMD’s MI430 series chips. These processors combine traditional supercomputing capabilities with AI application features. Oak Ridge National Laboratory, Hewlett Packard Enterprise, and AMD are designing this second system.

Discovery is scheduled for delivery in 2028 with full operations starting the following year. Streiffer anticipates substantial performance improvements but has not specified exact capabilities.

IBM announced separate progress using AMD chips for quantum computing. The company successfully ran quantum error-correction algorithms on standard AMD processors. The test achieved speeds ten times faster than baseline requirements for reliable operation.

Partnership Details and Market Response

The Department of Energy will host both supercomputers while private companies provide equipment and funding. Both parties will share access to the computing resources. A DOE official stated these AMD-based systems are intended as the first of multiple public-private partnerships across national laboratories.

Using AMD chips for quantum computing offers cost advantages and easier scalability according to IBM. The company plans to incorporate this approach into its Starling quantum system roadmap for 2029.

AMD stock rose 2.67% on Monday, closing at $259.67. The company has 39 analyst ratings with a Moderate Buy consensus. The average price target stands at $249.92. The supercomputer deal strengthens AMD’s position in government and research markets beyond its traditional gaming and cloud computing segments.

The post AMD Secures $1 Billion Energy Department Deal to Build Advanced Supercomputers appeared first on Blockonomi.