Europa Fusion is developing a different path toward fusion power: the startup plans to pair a pulsed-power reactor concept with a research strategy centered on simulations and data. Spotify founder Daniel Ek is backing the company, which was founded by AI researcher Nantas Nardelli.
A new company with an AI research background
Europa Fusion is based in the UK and was founded in February. Sources told TechCrunch that Nardelli had been working on the idea for longer, though they asked not to be identified because the information is not public. Bloomberg first reported Ek’s backing.
Nardelli has worked at Google DeepMind, Facebook AI Research, and Microsoft Research Cambridge. More recently, he was head of machine learning at Pacific Fusion, a fusion power startup based in Fremont, California.
That experience puts him at the intersection of two fields: machine learning and fusion research. Europa Fusion’s planned emphasis on simulation and data suggests that the company intends to bring methods from computational research into the work of developing a fusion power plant.
Why Europa Fusion is emphasizing simulations
Fusion startups often use physical devices to test their ideas and improve their simulations. Pacific Fusion, for example, leans heavily on hardware for that work. A source familiar with Nardelli’s thinking told TechCrunch that he did not agree with giving hardware that central a role, and believed fusion companies should invest more in simulation and data.
That difference concerns how a company learns as it develops its technology. Physical experiments can test theories directly, while simulations and data can help researchers investigate and refine them computationally. Europa Fusion is placing more weight on the latter approach, though the source does not describe the specific tools or processes the company plans to use.
The approach does not change the basic challenge: a startup still needs to develop a reactor concept that can produce useful energy. The source article offers no details about Europa Fusion’s facilities, schedule, or technical milestones, so its plans remain limited to the broad research strategy and reactor type it intends to pursue.
Pulsed power remains the reactor concept
Europa Fusion will use a pulsed-power approach, the same broad approach pursued by Pacific Fusion. In a pulsed-power reactor, large surges of electrical current arrive at precisely timed intervals and compress fuel targets. The compression brings atoms together so they fuse and release energy.
The concept relies on timing and powerful electrical pulses to create the conditions for fusion. Europa Fusion’s simulation-heavy strategy is therefore meant to support the development of a physical fusion system, rather than replace the need for one.
Pacific Fusion recently broke ground on a demonstration facility in New Mexico. The article gives no comparable construction update for Europa Fusion. Its reported distinction is the greater emphasis its founder wants to place on computational work as the company pursues pulsed-power fusion.
Fusion’s promise and the work ahead
Fusion could become a significant energy source in the coming decades, but no one has built a power plant capable of supplying electricity to the grid. That gap between potential and demonstrated grid power is central to the challenge for Europa Fusion and other companies working on the technology.
If fusion becomes commercially viable, it could provide carbon-free power and draw on nearly limitless fuel supplies. It could also help address concerns about the energy consumption of AI. Those are possible benefits, not outcomes the source says have already been achieved.
Nardelli has described climate change as a reason for moving from AI work toward fusion. On his website, he wrote: “My original background is in electronics and control engineering. I used to build robots for a living. One day I’ll be back to that, but it’s time to fix climate change for good.”
Europa Fusion’s plans bring together a specific reactor approach and a different idea about how to develop it. Whether a stronger reliance on simulations and data can help produce a working fusion power plant remains an open question, alongside the broader challenge of making fusion commercially viable.