Google DeepMind alumni found Fusionality to build fusion reactor control systems
Original: Google DeepMind alumni are building tools to accelerate fusion power for the grid
Why This Matters
Standardized fusion control infrastructure could significantly reduce development costs and timelines across the commercial fusion sector.
Former Google DeepMind researchers Federico Felici and Jonas Buchli founded Lausanne-based Fusionality in 2026, raising a $3.7M pre-seed round from Founderful and Playfair to develop standardized control systems and simulation tools for commercial fusion power companies.
Fusionality was founded by Federico Felici (CEO) and Jonas Buchli (CTO), both veterans of Google DeepMind's fusion research efforts. The startup is headquartered in Lausanne, Switzerland, and has raised $3.7 million (CHF 3 million) in pre-seed funding from Founderful and Playfair.
The founders identified a clear market gap: fusion companies are currently forced to build their own control systems from scratch, even though, according to Felici, approximately 80% of each company's control system is functionally identical. Fusionality aims to build a shared suite of control systems and simulation environments that individual fusion startups can then customize for their specific reactor designs.
Felici and Buchli previously collaborated on a landmark project teaching AI to control an experimental tokamak at EPFL in Switzerland, in partnership with Google DeepMind. Felici later joined DeepMind directly, where he developed simulations and machine learning interfaces for fusion devices.
On the role of AI in fusion control, Felici stated he would not advocate for AI to control an entire fusion reactor today, but noted it will "play an important role" in future systems by complementing, enhancing, or optimizing specific components. Fusionality joins a growing supply chain ecosystem serving the commercial fusion industry, alongside companies such as Kyoto Fusioneering.