Sometimes startup founders search high and low for the right business idea, while other times it arrives right on their doorstep. Federico Felici and Jonas Buchli are two of the lucky ones who found themselves uniquely positioned to solve a glaring problem in one of the world’s most complex and fast-moving technological sectors.
Over and over again, the two founders heard the exact same refrain from emerging nuclear fusion companies striving to commercialize clean energy. Industry players repeatedly noted that while they would ideally prefer to purchase many of the necessary components to build their proprietary control systems off-the-shelf, no such commercial providers existed. More importantly, they lacked suppliers who truly spoke the specialized language of nuclear fusion. Federico Felici highlighted this market gap, explaining that fusion developers constantly struggled to find reliable vendors who understood the intricate physics of superheated plasma and reactor management.
For years, both Felici and Buchli dedicated their professional careers to developing pioneering methods for controlling experimental fusion devices. They realized that a significant portion of their daily work involved solving precisely the types of complex engineering and software problems that commercial fusion startups now face as they race toward operational reactors. Recognizing the acute timing of the burgeoning industry, they decided it was the right moment to turn their expertise into a dedicated enterprise.
To capitalize on this immense market opportunity, Felici and Buchli founded Lausanne-based Fusionality earlier this year. The startup wasted little time in securing its financial foundation, swiftly closing a $3.7 million (CHF 3 million) pre-seed funding round backed by prominent venture capital firms Founderful and Playfair. Under the leadership of this founding team, Felici has taken the helm as the company’s chief executive officer, while Buchli drives technical development as the chief technology officer.
Fusionality now operates as a critical node in an emerging, highly specialized supply chain that has rapidly materialized to support the expanding global fusion power industry. The landscape is bifurcated into different types of specialized support players: some ambitious startups, such as Kyoto Fusioneering, focus on developing heavy engineering components that will help fusion companies translate their internal scientific breakthroughs into stable electricity delivered directly to the power grid. Others are traditional, high-precision manufacturing firms that possess the advanced fabrication skills required to build the extremely durable, tight-tolerance parts demanded by nuclear environments.
Yet, despite the growth of this supply chain, remarkably few companies specialize in designing the specialized hardware and software required to control the fusion reactors themselves. Across the broader industry, the standard practice remains remarkably inefficient, with many individual companies building their proprietary control systems entirely from scratch. Felici points out that this fragmented approach wastes valuable engineering hours because roughly 80% of each company’s control system architecture is actually identical across different reactor designs.
Nuclear fusion power plants are designed to generate vast amounts of clean electricity by harnessing the tremendous energy released when two atomic nuclei fuse together. This fundamental reaction is significantly more likely to occur when matter is brought to a superheated state known as plasma. However, managing plasma is notoriously difficult; these ionized gases are exceptionally fickle, and maintaining the proper internal temperature, magnetic containment shape, and continuous fuel level requires split-second, highly accurate decision-making from the reactor’s underlying infrastructure.
The control systems required to maintain these precise operational conditions are exceptionally complex to develop, test, and deploy in real-world settings. Nevertheless, much of the fundamental physics governing these reactions remains remarkably consistent across a wide variety of experimental reactor geometries. Fusionality plans to capitalize directly on this commonality by building a comprehensive suite of standardized control systems and sophisticated simulation environments. Fusion startups can then take these foundational tools and fine-tune them to seamlessly integrate with their specific reactor designs, saving precious time and development capital.
The origin of Fusionality traces back to the professional intersection of its founders. Felici and Buchli originally met while teaching advanced artificial intelligence to control an experimental tokamak, which is a complex, donut-shaped magnetic confinement device that researchers use to study the behavior of controlled fusion power. At the time, Felici was conducting research at EPFL in Switzerland, an institution that houses one such advanced tokamak facility. Meanwhile, Buchli was applying his expertise at Google DeepMind. Later, Felici also transitioned to Google DeepMind, where he spent his time developing sophisticated simulations and advanced machine learning interfaces specifically tailored for experimental fusion devices.
Reflecting on the role of modern software in this space, Felici noted that artificial intelligence will undoubtedly play an important role in future reactor control systems, though it is not yet fully prepared to tackle the entirety of the operational challenge alone. He emphasized that he would not advocate for algorithms to take over the total responsibility of controlling an entire nuclear fusion reactor in its current state of development. Instead, modern AI is far more likely to complement, enhance, or optimize specific subsystems within the broader architecture.
For its initial market entry, Fusionality is heavily focusing its engineering efforts on a prominent branch of fusion power known as magnetic confinement. This widely pursued approach relies on exceptionally powerful electromagnets to generate magnetic fields that keep fusion fuel dense enough and hot enough to sustain the reactions needed to feed a commercial power plant. Several high-profile startups are currently developing distinct magnetic confinement devices, including Commonwealth Fusion Systems, Realta Fusion, Proxima Fusion, and Type One Energy, among others. Felici noted that while magnetic confinement is the immediate priority, Fusionality eventually plans to expand its offerings to develop control systems for alternative fusion approaches as the market matures.
The newly secured pre-seed funding will allow Fusionality’s current team of seven employees to focus on developing a tightly selected group of core technologies. While Felici declined to publicly specify the exact nature of these upcoming products, he likened the company’s modular technology roadmap to a set of Lego blocks. The startup intends to start with a few foundational blocks and progressively build out a broader catalog, driven by the ultimate ambition to serve the remarkably wide range of technological subsystems required to safely and efficiently operate a commercial nuclear fusion reactor.