UKAEA and PPPL Agree to Explore Linking Fusion AI Supercomputers – Unite.AI

The United Kingdom Atomic Energy Authority (UKAEA) and the US Department of Energy’s Princeton Plasma Physics Laboratory (PPPL) signed a Joint Declaration of Intent on September 14, 2026, to explore linking UKAEA’s SUNRISE supercomputer with PPPL’s STELLAR-AI platform through a partnership known as the SUNRISE–STELLAR-AI Federation, the UKAEA announced.

The declaration was signed at the Global Fusion Policy Summit, held at Mansion House in London on September 14, 2026. The summit aims to increase regulatory harmonisation, which the announcement describes as crucial to achieving safe and secure fusion energy, and to create certainty for developers and investors. The release describes UKAEA and PPPL as the UK’s and one of the US’s leading national fusion laboratories.

The federation would connect the two platforms, enabling researchers on both sides of the Atlantic to train AI models on results from both national laboratories and to develop AI tools for future fusion power plants. Under the plan, both advanced computers would be used to train the same AI models, which the release notes require large volumes of high-quality data to make accurate predictions.

The Tokamaks and Supercomputers Behind the Federation

The collaboration would draw on experiments completed using UKAEA’s MAST Upgrade facility in Oxfordshire and PPPL’s NSTX-U facility in New Jersey. Both are compact spherical tokamaks with similar designs, which the announcement says makes them well suited for joint AI training.

Models built using data from a single machine can struggle to make accurate predictions that apply across different machines, according to the release. When different facilities are combined, researchers can develop more powerful machine-learning models that better capture the underlying physics, helping accelerate future fusion projects that include the UK’s Spherical Tokamak for Energy Production (STEP Fusion) and US concepts such as PPPL’s Spherical Tokamak Advanced Reactor (STAR).

The announcement describes SUNRISE as the UK’s first AI supercomputer dedicated to fusion energy, backed by £45 million from the UK government. It forms the first phase of computing infrastructure underpinning the UK’s first AI Growth Zone, established at UKAEA’s Culham Campus. STELLAR-AI is PPPL’s AI and high-performance computing platform, operated with support from Princeton University.

Aims and Official Statements

By combining supercomputing power and expertise, the federation aims to shorten the design cycle for future fusion power plants, make better use of existing experiments, and give the next generation of fusion scientists on both sides of the Atlantic a shared platform to build on, according to the announcement. The release frames the partnership as creating a world-leading transatlantic AI capability for fusion and as an investment in the people and skills needed for the future.

Joe Milnes, Executive Director for Engineering and Computing at UKAEA, described fusion as one of the major scientific and engineering challenges of the era and said meeting such challenges requires partnerships. He said the US and UK are two global leaders in fusion research and that federating SUNRISE and STELLAR-AI would build on a long history of transatlantic cooperation to further advance fusion development.

Jonathan Menard, Chief Scientist at PPPL, described a fusion power plant as one of the most complex machines humanity has attempted to build and said no single laboratory or nation will design one on its own. Federating the two platforms, he said, would let the partners train AI on data from two leading fusion facilities, test ideas across two of the world’s most capable computing systems designed for fusion research, and accelerate the path to a compact fusion power plant.

Rob Akers, Director of Computing Programmes and Senior Fellow at UKAEA, said the partners would draw on experimental data from both machines, using simulation to fill gaps and extend those datasets into operating regimes not yet explored. He said the teams would develop digital twins of both machines to support the design of future fusion power plants, producing models of spherical tokamaks that are more predictive and more useful to fusion engineers. He added that working as one team would let the partners learn faster and extract more insight from experiments.

Shantenu Jha, Head of Computational Sciences at PPPL, stated the technical goal: “Our goal is to let models and experiments move freely between the two systems. We will turn a collection of supercomputers into a single engine for fusion discovery.”

The June Agreement and What Comes Next

The federation builds on a memorandum of understanding between UKAEA and PPPL announced on June 25, 2026, which set out a framework for joint work between the two organizations. Under that agreement, the laboratories committed to expand scientific, academic and educational cooperation across fusion science and technology, including reciprocal staff exchanges, access to major research facilities, joint projects, the exchange of academic information, collaboration on ITER diagnostics, advanced computing programs and wider information sharing over the coming years.

The June agreement followed His Majesty The King’s Address to the Joint Meeting of Congress in Washington, D.C., which highlighted new UK-US partnerships on fusion energy, according to the June release. That release said the agreement would focus on advancing the physics and technology basis for future fusion power plants.

Following the summit, UKAEA and PPPL have discussed possible next steps for the federation, including exploring the development of digital twins of both facilities. The September release defines a digital twin as a detailed virtual copy of a real fusion machine, supported by regular data from fusion experiments, so researchers can test changes and predict how a machine will behave in software before applying experiments to hardware.

Researchers will also examine how to move computing jobs smoothly between the two platforms, which use different hardware, so scientists can pick the right supercomputing platform for each problem. The federation also plans to develop the next generation of fusion and AI experts through training, workshops, research opportunities and other international collaborations.