Two former Google DeepMind researchers, Federico Felici and Jonas Buchli, are turning years of fusion-control research into a new company aimed at making reactor systems more practical for the power grid.
The Lausanne-based startup, Fusionality, was founded this year after the pair repeatedly heard the same need from fusion developers: reliable control hardware and software built specifically for the demands of fusion. The company has already secured a $3.7 million pre-seed round from Founderful and Playfair.
Fusionality is focusing on a key gap in the emerging fusion supply chain. While many companies are building reactor components, fewer are developing the control systems that keep a plasma stable by managing temperature, shape, and fuel in real time. The startup believes much of that technology can be standardized and adapted across different reactor designs.
Felici and Buchli first worked together on AI systems for an experimental tokamak at EPFL in Switzerland. Their later work at Google DeepMind included simulations and machine learning tools for fusion devices, giving them a strong technical base for the new venture.
The company's first phase centers on magnetic confinement, the approach used by several well-funded fusion startups. Fusionality plans to build a set of control and simulation tools that reactor teams can customize rather than create from scratch.
With a seven-person team and a focused roadmap, Fusionality is betting that shared infrastructure can help fusion move faster from laboratory progress toward grid-ready energy. If successful, such tools could help shape the next era of clean power systems.
