Two former researchers from Google DeepMind have launched Fusionality, a new startup aiming to standardize the control systems used in experimental fusion reactors. Co-founders Federico Felici and Jonas Buchli raised a $3.7 million (CHF 3 million) pre-seed round from investors Founderful and Playfair.
The company is based in Lausanne, Switzerland, and addresses a critical gap in the emerging fusion supply chain. While some firms focus on manufacturing components or developing reactor designs, few specialize in the integrated hardware and software required to manage fusion operations.
“Over and over again, they heard the same thing from fusion companies: ‘We would ideally buy many of the components to make our control system, but there isn’t really anybody providing them — and especially there isn’t anybody who speaks the language of fusion,'” said Felici.
Felici and Buchli met while working on AI-driven control for tokamaks, the donut-shaped devices that use magnetic fields to confine plasma. Their prior experience at EPFL in Switzerland and Google DeepMind, where they developed simulations and machine learning interfaces for fusion devices, provided the foundation for the new venture.
According to Felici, approximately 80% of the control systems across different fusion startups are fundamentally identical, yet each company typically builds theirs from scratch. Fusionality plans to leverage this commonality by creating a suite of modular control systems and simulation environments that can be customized for specific reactor designs.
For now, the startup is focusing on magnetic confinement fusion, the approach utilized by companies such as Commonwealth Fusion Systems, Realta Fusion, Proxima Fusion, and Type One Energy. Felici noted that while AI will play an important role in future control architectures, it is currently best suited to complement and optimize specific parts of the system rather than manage the entire reactor autonomously.
Fusionality currently employs seven people. The pre-seed funding will support the development of a “tightly selected number of technologies,” which Felici compared to Lego blocks that can be expanded upon over time. He expressed the ambition to eventually serve the full range of technologies required to operate commercial fusion power plants.
I’m skeptical about leveraging off-the-shelf modules for such complex physics. Fusion varies wildly between designs; can Lego blocks really work?
Standardizing control systems is a brilliant move. Building from scratch wastes so much effort when 80% is identical!