The United States Department of Energy has entered into a $1 billion partnership with Advanced Micro Devices (AMD) to develop and build two new artificial intelligence supercomputers. The initiative aims to address complex scientific challenges, including fusion energy, advanced cancer treatments, and national security.
The collaboration will produce two systems, named Lux and Discovery, which will be hosted at Department of Energy facilities. The project represents a new model of public-private partnership intended to accelerate scientific breakthroughs using next-generation computing power.
Key Takeaways
- The U.S. Department of Energy and AMD have announced a $1 billion partnership for AI supercomputing.
- Two supercomputers, Lux and Discovery, will be built to tackle major scientific problems.
- Lux is scheduled to come online within six months, using AMD's MI355X AI chips.
- Discovery, a more advanced system using MI430 chips, is expected to be operational by 2029.
- The initiative aims to accelerate research in fusion energy, cancer treatment, and national security.
A New Era of Scientific Computation
Government officials and corporate leaders have unveiled a major investment designed to maintain the nation's leadership in high-performance computing. The partnership leverages capital and technology from the private sector to build infrastructure for public scientific research.
Energy Secretary Chris Wright stated that the new systems will "supercharge" progress in critical areas. He highlighted the potential for significant advancements in fusion energy, a clean energy source that scientists have pursued for decades.
"We've made great progress, but plasmas are unstable, and we need to recreate the center of the sun on Earth," Wright said. "We're going to get just massively faster progress using the computation from these AI systems that I believe will have practical pathways to harness fusion energy in the next two or three years."
Beyond energy, the project has ambitious goals in medicine. Wright expressed hope that the computational power could revolutionize cancer treatment by simulating molecular-level therapies.
"My hope is in the next five or eight years, we will turn most cancers, many of which today are ultimate death sentences, into manageable conditions," he added.
National Security Implications
In addition to scientific research, the supercomputers will play a role in national security. Officials confirmed the systems would be used to help manage the U.S. nuclear weapons arsenal, a task that requires immense data-processing capabilities for simulations and maintenance.
Lux: Speed and Immediate Impact
The first of the two machines, named Lux, is on an accelerated timeline. It is expected to be constructed and brought online within the next six months. This rapid deployment is a key feature of the project.
AMD CEO Lisa Su commented on the speed of the project, noting, "This is the fastest deployment of this size of computer that she has seen." She emphasized that this agility is crucial for advancing U.S. AI efforts.
The Technology Powering Lux
Lux will be built using a suite of AMD technologies, creating a highly integrated system. The core components include:
- AMD Instinct MI355X AI accelerators: These chips are designed specifically for large-scale AI and high-performance computing tasks.
- AMD CPUs: The system will also utilize AMD's central processing units for general computation.
- AMD Networking Chips: Specialized networking hardware will ensure rapid data transfer between components.
The development of Lux is a collaborative effort involving AMD, Hewlett Packard Enterprise (HPE), Oracle Cloud Infrastructure, and the Oak Ridge National Laboratory (ORNL).
According to ORNL Director Stephen Streiffer, the Lux supercomputer is projected to deliver approximately three times the AI capacity of the most powerful supercomputers currently in operation.
Discovery: The Future of High-Performance Computing
Following Lux will be a more advanced system named Discovery. This second supercomputer is slated for delivery in 2028 and is expected to be fully operational in 2029.
Discovery will be based on AMD's next-generation MI430 series of AI chips. According to Lisa Su, the MI430 is a special variant of the MI400 family, uniquely engineered to combine features of traditional supercomputing chips with the capabilities needed for modern AI applications.
This hybrid design is intended to provide massive performance gains for a wide range of scientific simulations and data analysis tasks. While specific performance projections were not provided, Stephen Streiffer said he expected "enormous gains" in computational capability from the new architecture.
A Public-Private Partnership Model
The financial and operational structure of this initiative marks a strategic shift. A Department of Energy official explained that the DOE will host the computers at its national labs, while the corporate partners—AMD, HPE, and Oracle—will provide the machines and associated capital spending.
In return for their investment, both the government and the private companies will share the immense computing power of the systems. This model is intended to be the first of many similar partnerships across DOE labs in the country.
By combining federal oversight with private sector agility and capital, the U.S. aims to ensure it has sufficient supercomputing resources to handle increasingly complex experiments and maintain its competitive edge in science and technology.