Two fusion supercomputers in the UK and the US can be connected to train identical AI models on experimental data from separate fusion machines.
The collaboration could help engineers design future power plants faster. The proposed Sunrise-Stellar-AI federation links the UK Atomic Energy Authority’s Sunrise supercomputer with the STELLAR-AI platform at Princeton Plasma Physics Laboratory.
Researchers will use data from the MAST Upgrade facility in Oxfordshire and the NSTX-U facility in New Jersey. The two facilities are compact spherical tokamaks with similar designs, and this similarity could help researchers combine experimental results and train models that better predict fusion plasma behavior across different machines.
The partnership was announced at the Global Fusion Policy Summit in London on 14 September 2026.
Joe Milnes, executive director of engineering and computing at the UK Atomic Energy Authority, says: “Fusion is one of the greatest scientific and engineering challenges of our time. Solving these challenges requires collaboration in fusion.
The US and the UK are two global leaders in fusion research, and the Sunrise and Stellar-AI federation will build on a long history of transatlantic collaboration.”
According to ITNA and quoting Interesting Engineering, Sunrise is the UK’s first AI supercomputer dedicated to fusion energy and is supported by £45 million of government funding. Stellar-AI provides AI and high-performance computing capabilities at Princeton Laboratory.
The planned system is also intended to support the development of digital twins for the two tokamak facilities.
A digital twin creates an accurate virtual model of a fusion machine using experimental data, and researchers can test changes in software and predict machine behavior before attempting them on physical equipment.