Google DeepMind alumni are building tools to accelerate fusion power for the grid

Two former Google DeepMind engineers, Federico Felici and Jonas Buchli, have founded Fusionality to address a critical gap in the burgeoning fusion power industry: specialized control systems. Observing that many fusion startups "build their own control systems from scratch," despite significant overlap in core requirements, Fusionality aims to provide modular hardware and software "Lego blocks" that can be customized for various magnetic confinement fusion devices. The company secured $3.7 million in pre-seed funding from Founderful and Playfair to develop these essential components. This development is crucial as it accelerates the commercialization timeline for fusion power by reducing development overhead for individual companies, allowing them to focus on core fusion breakthroughs rather than foundational control infrastructure. The venture taps into the growing need for a specialized supply chain to enable fusion energy's transition from experimental physics to grid-scale power generation, with AI poised to play a complementary role in optimizing these systems.

AI Signal Decode

Fusionality's core value proposition lies in standardizing and modularizing the complex control systems required for fusion reactors. Founders identified that a significant portion, approximately 80%, of control system development is common across different magnetic confinement fusion designs. By offering pre-built, customizable "Lego blocks" of hardware and software, the company enables startups to avoid the costly and time-consuming process of building these systems from the ground up. This approach directly addresses a bottleneck in the fusion industry, accelerating development cycles and potentially reducing the cost of bringing fusion power to the grid.

The market implication of Fusionality's offering is substantial. The fusion industry is characterized by numerous startups pursuing different technological paths, all requiring robust and precise control mechanisms to manage superheated plasma. By providing a specialized supply chain solution, Fusionality lowers the barrier to entry and accelerates the commercialization roadmap for magnetic confinement fusion companies like Commonwealth Fusion Systems and others. The $3.7 million pre-seed funding will be instrumental in developing the initial set of technologies, positioning Fusionality as a key enabler within this rapidly expanding sector.

Technically, Fusionality leverages the founders' deep expertise in AI and fusion control systems, gained at Google DeepMind and EPFL. While acknowledging AI's future role in enhancing and optimizing control, the company's immediate focus is on foundational control systems for magnetic confinement. This pragmatic approach, focusing on established, albeit complex, control paradigms, allows for near-term impact. Future development will likely expand to other fusion approaches, indicating a long-term vision to become a comprehensive provider of control solutions across the diverse fusion landscape.

Looking ahead, it will be critical to observe Fusionality's progress in developing and deploying its initial "Lego blocks" of control technologies. Successful adoption by magnetic confinement fusion startups will validate their modular approach and market demand. Furthermore, the company's ability to adapt its systems for different reactor designs and potentially for other fusion methodologies will be key to its long-term success and its impact on the broader goal of achieving grid-scale fusion power.