Fusionality has secured CHF 3 million (approximately $3.7 million) in pre-seed funding, with Founderful and Playfair leading this round.
The founders had worked at Google DeepMind before starting Fusionality, conducting research on AI to control EPFL’s tokamak.
Over 30 private fusion companies currently require the plasma software that Fusionality is developing.
Federico Felici and Jonas Buchli have, for many years, developed AI to prevent a plasma cloud at a temperature of 100 million degrees from coming into contact with the reactor’s walls. They now wish to share this knowledge.
The two have secured CHF 3 million, which is approximately $3.7 million, in pre-seed funding, led by Founderful and Playfair, for Fusionality, the startup they run in Lausanne that produces measurement, simulation, and control systems for fusion companies.
“We have spent most of our careers making plasmas in real fusion devices behave as we wanted. Now we founded Fusionality to bring that expertise to our customers and partners, helping them tackle some of the hardest bottlenecks in fusion operations,” said Felici, co-founder and chief executive of Fusionality.
More than thirty private fusion companies and government projects face the same problem: converting raw plasma measurements into signals that control the heating, fuelling, and magnetic systems within milliseconds.
Private fusion investment reached a record $4.48 billion over the twelve months ending July 2026, the Fusion Industry Association states. Since 2021, the sector has seen over $14 billion in total funding distributed among 56 companies.
However, most of this money is not invested in reusable infrastructure, since each company generally designs its own control systems from scratch.
How a research paper in Nature led to a startup
Felici and Buchli came together at EPFL’s Swiss Plasma Center. Felici, a control engineer with a PhD in plasma physics, was involved with the TCV tokamak before spending two and a half years at Google DeepMind. Buchli, an electrical engineer from ETH Zürich, headed the robotics and reinforcement learning teams at DeepMind.
They and their colleagues applied deep reinforcement learning to control the magnets that shape the plasma in the TCV and published their findings in Nature in 2022.
Buchli, who is now the chief technology officer at Fusionality, stated that operating a fusion device involves dealing with multiple sources of data, different time scales and complex plasma behaviour, and that Fusionality features a multi-disciplinary approach covering measurement, software, control and data engineering.
Fusionality has a basic software layer that requires minimal adaptation, helping customers reduce their onboarding costs. In this way, it competes with Next Step Fusion, a company in the field of supply-chain management that already supplies plasma modelling, diagnostics, and control software to tokamak builders.
Google is also facing a challenge in this situation. In October 2025, DeepMind entered into a partnership with Commonwealth Fusion Systems, made the TORAX plasma simulator available as an open-source tool, and applied reinforcement learning to CFS’s SPARC reactor. Since the most well-funded fusion company can obtain DeepMind’s expertise at no cost, the market for paid third-party solutions may decrease.
The money, the voices, and what’s next
Felix Neubeck from Playfair stated, “The fusion industry is scaling faster than its supply chain, and the operations infrastructure is the clearest gap, Federico and Jonas are among a very small number of people in the world who have worked across the range of technologies needed for fusion control & operations, on various devices. That combination of academic standing and commercial instinct is rare, and it is why we backed them.”
Paolo Ricci, the director of the Swiss Plasma Centre, added, “The Swiss Plasma Center at EPFL has decades of experience in plasma control and diagnostics. It is a great development for the field, and for the Swiss fusion ecosystem, to witness the creation of startups like Fusionality, benefiting from this knowledge and serving the emerging fusion industry worldwide.”
The money will be used to employ control engineers, computational physicists, plasma measurement specialists and software engineers in Lausanne, as well as to secure Fusionality’s first paying customers. Since its founding, Founderful has supported the EPFL spin-out Isospec Analytics, and Playfair has also invested in the UK fusion-isotope startup Astral Systems.
Whether Fusionaluty is adopted as the common infrastructure by an industry consisting of 56 companies, or whether it is displaced by reactor builders carrying it out themselves, with a well-funded laboratory in London providing free solutions, will show more about how commercial fusion will be developed than any single achievement of a reactor.