DOE and AMD pledge USD 1 billion for two AMD‑chip AI supercomputers at Oak Ridge
When the Department of Energy and AMD announced a roughly $1 billion pledge, the headline was obvious: Oak Ridge National Laboratory is about to get a serious upgrade. The money will pay for two brand-new systems that run on AMD’s latest CPUs and GPUs, basically extending the same architecture that powers the Frontier supercomputer. It looks like the government is trying to funnel more cash into AI-focused hardware, probably to keep the U.S.
a step ahead in big-scale machine-learning research and scientific simulation. Some insiders say it’s a strategic push to grow home-grown compute power, especially now that AMD has rolled out a handful of notable GPU and CPU improvements. At the same time, the deal hints at a wider pattern of public-private partnerships aimed at the next wave of AI workloads.
With the agreement now out in the open, the next move will be a formal announcement that strings together the agencies, the exact investment figure and the hoped-for impact on the country’s AI infrastructure.
AMD, Department of Energy announce $1 billion AI supercomputer partnership Oak Ridge National Laboratory is getting two new supercomputers powered by AMD chips. Oak Ridge National Laboratory is getting two new supercomputers powered by AMD chips. Both build on the work that went into the Frontier supercomputer, which is also housed at ORNL and was the fastest in the world until El Capitan came online last year at Lawrence Livermore National Laboratory. AMD also helped develop those supercomputers, so this isn't its first time working with the US government on a project like this.
The $1 billion price tag on the new machines raises a lot of questions. AMD and the Department of Energy have pledged that sum to build two AI-focused supercomputers, Lux, which should appear in early 2026, and Discovery, slated for 2029, both destined for Oak Ridge National Laboratory. Oracle and Hewlett Packard Enterprise are in the mix, and the systems will run on AMD’s newest silicon, an evolution of the tech that drove the Frontier machine.
Lux will arrive relatively soon, giving researchers a chance to try the hardware on real-world jobs; Discovery’s later timeline seems to allow more design tweaks. The article, however, leaves out performance goals, power use, or how the rigs will fit with other DOE projects. It’s unclear whether the partnership will push AI research forward or simply add more nodes to an already busy field.
As the rollout dates draw nearer, I expect the scientific community to focus on actual results rather than just the size of the investment.
Common Questions Answered
What is the total amount of funding the U.S. Department of Energy and AMD have pledged for the new AI supercomputers at Oak Ridge National Laboratory?
The partnership commits a combined $1 billion to develop two AI‑focused supercomputers at Oak Ridge. This funding is intended to cover hardware acquisition, system integration, and related research infrastructure over the next several years.
Which two AI‑focused supercomputers are planned for Oak Ridge, and when are they expected to become operational?
The two systems are named Lux and Discovery. Lux is slated for early 2026 deployment, while Discovery is projected to be operational by 2029.
How will the architecture of Lux and Discovery build on the design of the Frontier supercomputer?
Both Lux and Discovery will extend the architecture that powered Frontier, utilizing AMD’s latest silicon and design principles. The new machines will incorporate advanced CPUs and GPUs that improve AI workloads while retaining the high‑performance computing capabilities demonstrated by Frontier.
Which other technology partners are involved in the construction of the new Oak Ridge supercomputers, and what are their contributions?
Oracle and Hewlett Packard Enterprise are key partners in the projects. Oracle provides cloud and data‑management services, while Hewlett Packard Enterprise supplies the chassis, networking, and system‑integration expertise needed to assemble the supercomputers.