Energy

Google and RWE Back German Fusion With 411 Million

Munich fusion startup Proxima Fusion raised 411 million euros at a 2.4 billion valuation, with Google and the utility RWE as strategic backers. When a data-centre power buyer and a grid operator co-fund a plant a decade out, they are pre-buying firm power before it exists, and a retired Bavarian reactor site is where it will be built.

EnergyBy Servola Tech Desk2026-07-083 min read

AI-assisted, edited by humans. | Editor-in-Chief: Leon Soliman | Editorial standards

Google and RWE Back German Fusion With 411 Million

Key takeaways

A hyperscaler and a utility just put 411 million euros into a Munich fusion startup at a 2.4 billion euro valuation, which tells you the buyers of firm clean power are now pre-funding it a decade before a single kilowatt-hour exists.
Google made its first European fusion investment and the German utility RWE came in as a strategic backer months after agreeing to host Proxima's first stellarator plant on the site of the retired Gundremmingen nuclear station in Bavaria, so the demand side and the grid side are both already committed.
The scarce asset in this deal is not the reactor physics but the grid connection: a decommissioned nuclear site brings an existing high-voltage interconnection, cooling and a licensing history, which in a market where new grid links can take seven to ten years is worth more than the technology bolted onto it.

What Proxima raised, and from whom

Proxima Fusion, a Munich company spun out of the Max Planck Institute for Plasma Physics, has raised 411 million euros, about 468 million dollars, at a valuation of 2.4 billion euros. The round was led by the investment firms XTX Ventures and East X Ventures, with the search and cloud group Google and the German energy utility RWE joining as strategic investors. For Google it is a first fusion investment in Europe, and the company said the money will expand its production of high-temperature superconducting cable and magnets and build out the engineering needed for its stellarator design.

The number matters less than the identity of the two strategic names on the cap table. XTX and East X are there for the return; Google and RWE are there for the electricity. A cloud operator whose growth is now gated by power availability and a utility that has to keep a national grid supplied are not passive financial investors in a science project. They are the eventual customer and the eventual host, buying a seat at the table of a technology that, if it works, produces exactly the firm, weather-independent, carbon-free power both of them are short of.

Why a utility and a hyperscaler are the tell

Proxima is not promising power soon. Its plan runs through a net-energy demonstrator targeted for the early 2030s and a first commercial stellarator plant, called Stellaris, later that decade. That timeline is the point, not a caveat. When a hyperscaler and a utility commit capital to generation that is ten years out, they are signalling that they expect firm clean power to be scarce and expensive for the whole of that decade, and that pre-positioning now is cheaper than competing for it later. This is procurement dressed as venture capital.

The logic is the same one already visible in the small-reactor and long-duration-storage deals hyperscalers have signed: the constraint on data-centre and industrial growth has moved from chips and capital to megawatts that are available around the clock. A stake in Proxima does not solve that this year, but it gives Google an early claim on a future firm-power source and gives RWE a next-generation asset to put on ground it already controls. Owners in any power-hungry business should read the co-investment as a forecast: the two parties closest to the meter are betting the squeeze lasts.

The retired reactor site is the real asset

The most revealing detail is where the first plant would go. RWE became an investor months after agreeing to partner with Proxima on building its first stellarator plant at Gundremmingen in Bavaria, the site of a nuclear fission station that has been shut down. That choice is not sentimental. A decommissioned nuclear site arrives with a high-voltage grid connection already built, cooling and water infrastructure in place, land zoned for power generation, and a community and regulator already used to a plant on that ground. Those are precisely the assets that take years to assemble from scratch.

In a market where a fresh grid connection can wait seven to ten years, an existing point of interconnection is worth more than the machine attached to it. The Gundremmingen plan says the real land bank for next-generation generation, whether fusion or small modular reactors, is the map of retired and retiring nuclear and coal sites, because they already hold the scarce permits and wires. Operators planning long-horizon energy supply should be looking at that map now, since the grid connection, not the reactor, is the part that cannot be bought quickly.

Frequently asked questions

How much did Proxima Fusion raise, and at what valuation?

Proxima Fusion raised 411 million euros, roughly 468 million dollars, at a valuation of 2.4 billion euros, about 2.7 billion dollars. The round was led by XTX Ventures and East X Ventures, with Google and the utility RWE joining as strategic investors. It is one of the best-funded fusion companies in Europe.

When will Proxima actually produce electricity?

Not for years. Proxima is targeting a net-energy demonstrator in the early 2030s and a first commercial stellarator plant, Stellaris, later in that decade. The investment is a long-horizon bet on firm clean power, not a near-term supply. That distance is why the involvement of a utility and a hyperscaler is the notable part.

Why does it matter that Google and RWE are the backers?

Because they are the demand side and the grid side of the power market, not passive investors. Google buys enormous amounts of round-the-clock electricity for its data centres, and RWE runs generation and grid assets. Their co-investment is a signal that the buyers and hosts of firm clean power expect it to stay scarce for a decade and are pre-positioning now.

Strip away the fusion headline and this is a procurement decision. The company that buys power by the gigawatt and the utility that has to deliver it both put money into generation that will not run until the 2030s, and they chose to build it on a retired reactor site that already has the wires. Firm clean power is scarce enough that the parties closest to the meter are pre-buying it a decade early, and the grid connection under an old nuclear plant has quietly become the asset worth owning.

Proxima FusionFusion EnergyGoogleRWEStellaratorGrid Connection

CMA Moves to Pry Open Apple and Google

The UK CMA has opened consultations to force Apple and Google to allow off-platform payment steering and third-party NFC access. What owners and operators need to weigh now.

3 min read

Law & Regulation

Google's 4.1 Billion Android Fine Is Final

The EU's top court has upheld Google's 4.1 billion euro Android fine with no further appeal. Here is what the final ruling means for platform owners.

2 min read

Digital Sovereignty

The EU Council Revived the Rule That Scans Private Messages

The EU Council revived the lapsed rule that lets providers scan private messages with AI. For owners, confidential communication is now a governance choice.