Summary

• Proxima Fusion has raised €411m from investors including Google, RWE, XTX Ventures, and East X Ventures.
• The Munich company is developing stellarator based fusion technology and working with RWE on a future Bavarian demonstrator.
• The investment reflects how large compute users are searching for long term clean power options, even in technologies that remain commercially unproven.

Proxima Fusion has raised €411m from investors including Google and RWE, drawing one of Europe’s most ambitious fusion companies into the energy contest now surrounding AI infrastructure.

The Munich based company said the round was led by XTX Ventures and East X Ventures, with Google and German utility RWE joining as strategic investors. The financing values Proxima at about €2.4bn, while RWE has invested €25m and signed an agreement around building a stellarator fusion power plant at the site of a former nuclear fission facility in Gundremmingen, Bavaria.

Fusion remains a long term technology. No private company has yet proved commercial electricity generation from fusion at grid scale, and the sector is full of technical, financial, regulatory, and manufacturing risk. Proxima’s approach is based on stellarators, a form of magnetic confinement system that aims to hold superheated plasma in a stable configuration long enough to produce useful power.

The financing stands out because of the type of capital now entering the field. Google’s participation links the search for future clean power to the growth of AI infrastructure. XTX, a trading firm with heavy compute requirements, has also been building datacentre capacity in Finland. RWE brings utility expertise and potential project development capability. Together, those investors suggest fusion is being evaluated not only as climate technology, but as future infrastructure for power hungry digital systems.

That does not mean fusion is about to solve the datacentre electricity problem. AI facilities are being built now, grid queues already stretch across many markets, and power demand is rising faster than experimental energy technologies can be commercialised. Even optimistic fusion timelines point to the 2030s for demonstrators and later for commercial plants. The mismatch between immediate demand and speculative future supply should temper any claim that technology companies can invest their way out of near term grid constraints.

Even so, the investment reveals how AI economics are changing. Compute is constrained not only by chips, models, and software talent, but also by land, interconnection, cooling, electricity procurement, and the credibility of clean power commitments. Cloud and AI companies are spreading bets across renewables, nuclear fission, geothermal, storage, demand management, and fusion because future compute growth depends on reliable energy that can survive policy scrutiny.

Europe has a particular interest in keeping that capital attached to domestic industrial capability. Proxima spun out of the Max Planck Institute for Plasma Physics and has built partnerships involving Bavaria, RWE, and the Max Planck ecosystem. The company’s plans include a demonstration reactor and a wider Bavarian project designed to turn research expertise into industrial advantage.

The risk is that fusion becomes another field where European science creates globally valuable intellectual property while commercial scale is captured elsewhere. If Proxima can use the financing to build supply chains, magnet manufacturing, engineering teams, and utility partnerships inside Europe, the upside extends beyond a single clean energy bet. If not, the round risks becoming another example of strategic capital flowing around European deep tech without resolving the continent’s scale up and deployment weaknesses.

For now, the Proxima round is a signal of demand rather than proof of delivery. AI infrastructure has made clean, firm power a boardroom issue for technology companies, and fusion offers a tempting answer because it promises abundant electricity without the land footprint of renewables or the waste politics of fission. Whether it becomes usable infrastructure will depend on engineering reality, not investor appetite.