Google Backs €411M Round for German Fusion Startup Proxima, Betting on 2030s Power Plant

When Google announced a fusion energy investment in Germany on July 7, it was easy to read the move as the kind of speculative, moonshot-adjacent bet that big tech companies make to burnish their sustainability credentials. Look at the numbers, however, and the investment logic becomes much harder, much more commercial, and much more urgent.

German startup Proxima Fusion has closed a €411 million ($469 million) Series C at a €2.4 billion ($2.7 billion) valuation, making it the most valuable fusion company in Europe by a substantial margin. The round was led by XTX Ventures—the investment arm of algorithmic trading firm XTX Markets, whose founder Alex Gerko has quietly become one of the most active deep-tech investors in the world—alongside London-based East X Ventures. Google (through Alphabet) and German national energy firm RWE AG are strategic investors in the round.

Proxima’s goal: build Europe’s first commercial fusion power plant, with a demonstrator operating in the early 2030s.

Why Stellarators, and Why Now

Proxima is developing a type of fusion reactor called a stellarator—a cousin of the more famous tokamak design but with fundamentally different magnetic field geometry. Where tokamaks (the design used in ITER and by Commonwealth Fusion Systems) require precise plasma control that is still an active research challenge, stellarators are theoretically more stable because their twisted magnetic fields can confine plasma without the constant active corrections that tokamaks need.

The world’s most advanced stellarator, Wendelstein 7-X at Germany’s Max Planck Institute for Plasma Physics in Greifswald, demonstrated key milestones in recent years that gave Proxima’s founders—many of whom came directly from the Max Planck team—the confidence that commercial stellarators were achievable within a decade. The company was founded in 2023, which means it has gone from inception to a $2.7 billion valuation in three years, a pace that reflects both the genuine technical progress and the extraordinary urgency of the energy problem the company is attempting to solve.

The AI Energy Crisis Behind the Investment

Google’s participation in this round cannot be understood without understanding Google’s energy problem.

Google’s AI operations consumed 37% more electricity in 2025 than the year before—an acceleration that has made the company the largest single purchaser of renewable energy in the world but also pushed it toward urgent exploration of power sources that renewable energy alone cannot provide. Solar and wind are intermittent. Batteries at the scale Google needs don’t exist. Nuclear fission provides baseload, but permitting a new fission plant in Europe can take fifteen to twenty years.

Fusion promises something fission can’t easily deliver: relatively compact plants with no long-lived radioactive waste, no proliferation concerns from the fuel cycle, and potentially fast deployment timelines once the technology is proven. Google has already signed an offtake agreement with Commonwealth Fusion Systems (CFS) in the United States, committing to buy electricity from CFS’s first commercial plant once it is operational. The Proxima investment extends that hedge across the Atlantic, covering European data center loads.

The strategic logic is straightforward. Google needs to know that reliable, carbon-free baseload electricity will be available in Europe in the 2030s, when its AI infrastructure requirements will be far larger than today. Investing in Proxima buys optionality: if Proxima hits its milestones, Google has a credible path to near-zero-carbon baseload in Germany and potentially beyond. If Proxima misses, Google has lost a relatively small amount of money relative to its overall capital position while gaining deep visibility into the state of European fusion development.

RWE’s Strategic Calculus

The participation of RWE AG is separately significant and arguably underappreciated in the initial coverage of the round.

RWE is not a venture capital firm making speculative bets. It is Germany’s largest electricity generator, a company that survived the transition away from coal and nuclear fission and rebuilt itself as a major operator of offshore wind, solar, and gas-fired peaking capacity. Its investment in Proxima is a bet from the demand side: a utility that will need to replace baseload capacity in the 2030s and 2040s, as Germany’s current energy mix leaves significant gaps that wind and solar cannot fill on their own.

RWE’s involvement also provides Proxima with something that pure financial investors cannot: a credible downstream customer with existing grid connections, regulatory relationships, and the operational experience to actually integrate fusion power into the German grid once a plant is running. The relationship between a fusion startup and a utility partner is not merely financial—it is the critical bridge between a technology demonstration and a commercial product.

The European Fusion Race

Proxima is not alone. Across Europe, a small number of well-capitalized fusion startups are competing to be first to commercial power. British startup Tokamak Energy is pursuing a more compact spherical tokamak. Gauss Fusion in Germany, backed by EWE (the northern German utility), is working on a conventional ITER-derived approach. French startup Renaissance Fusion is developing a high-temperature superconducting magnet design for a compact tokamak.

What distinguishes Proxima is the stellarator approach—which, if it works at scale, could offer inherently better plasma stability than the alternatives—and the direct lineage from the Wendelstein 7-X team, which gives the company a credible technical foundation that pure venture-backed startups often lack.

The €411 million round is a record for European fusion. For context, the entire European fusion startup sector raised roughly €1.5 billion between 2020 and 2024 across all companies. Proxima has now raised more in a single round than most European fusion competitors have raised in their entire existence.

The Timeline Problem

The central challenge for fusion investment—and the reason it has attracted justified skepticism for decades—is that fusion has always been “twenty years away.” Proxima’s timeline targets a demonstrator in the early 2030s and a commercial plant by the mid-to-late 2030s. That is aggressive, and there are serious technical hurdles remaining.

Plasma confinement at commercial scale in a stellarator has never been demonstrated. The engineering challenges of building a system that can withstand the neutron bombardment from sustained fusion reactions are formidable. The materials science for plasma-facing components at commercial scale does not fully exist yet. And the grid integration of a new type of power plant—even a well-functioning one—introduces regulatory, permitting, and operational challenges that can easily add years to any timeline.

None of this means Proxima’s projections are wrong. Commonwealth Fusion Systems and TAE Technologies have made similar aggressive projections, and the technical community’s view of the feasibility of near-term fusion has shifted meaningfully as high-temperature superconducting magnets have made dramatic progress. But investors putting €411 million into a company expecting a commercial power plant in twelve years should understand they are making a bet on both the technology and the regulatory environment—and that both can move in unpredictable ways.

What This Means for the AI Industry

For the tech industry specifically, the Google-Proxima relationship represents a significant development in how AI companies are thinking about energy security. The era of assuming that the electricity grid will simply absorb AI’s demand growth is over. The hyperscalers are now direct participants in the energy supply chain, signing offtake agreements with nuclear fusion startups, building their own gas plants, and lobbying for accelerated permitting of fission plants in the United States and Europe.

The question that will define energy policy for the next fifteen years is not whether AI needs more power—it clearly does—but whether the infrastructure to supply that power can be built quickly enough to keep pace. Google’s bet on Proxima is, in its essence, a vote that it cannot rely on existing technology alone to solve that problem, and that fusion—despite its historically frustrating timeline—is worth funding seriously.

That is a significant signal from one of the most analytically rigorous capital allocators in the technology sector.