Google just put real money behind one of the hardest engineering problems humanity has ever attempted. On July 7, 2026, the company backed German fusion startup Proxima Fusion in a €411 million (about $468 million) funding round — the largest private fusion investment in Europe's history.
It's not Google's first fusion bet, and it won't be the last from Big Tech. Microsoft, Amazon, and OpenAI's Sam Altman have all made similar moves in the past two years. The reason is simple: AI is eating electricity faster than the grid can supply it, and fusion promises something the industry hasn't had in a century — a genuinely new source of clean, dense power.
Here's what actually happened, what fusion power really is, and why it's suddenly a boardroom priority instead of a physics-department curiosity.
What Google Actually Invested In
Proxima Fusion is a three-year-old startup spun out of Germany's Max Planck Institute for Plasma Physics. The €411 million round was led by XTX Ventures and East X Ventures, with Google and German energy giant RWE joining as strategic investors. The deal values Proxima at roughly €2.4 billion (about $2.7 billion), making it Europe's best-funded fusion company by a wide margin.
The money will fund a project called Alpha — a "net-energy stellarator demonstrator" being built near Munich. Proxima is aiming to have Alpha running in the early 2030s, as a proof-of-concept step toward a full commercial power plant it calls Stellaris, planned for later this decade at the site of a former nuclear fission plant in Gundremmingen, Bavaria.
That detail is worth sitting with for a second: a facility that once split atoms for power could end up fusing them instead.
Fusion vs. Fission: Why the Distinction Matters
Every nuclear power plant running today uses fission — splitting heavy atoms like uranium apart to release energy. It works, but it produces long-lived radioactive waste and carries meltdown risk if something goes wrong.
Fusion works the opposite way: it forces light atoms, usually hydrogen isotopes, to combine into helium — the same reaction that powers the sun. Done right, fusion offers a few major advantages:
- Abundant fuel — hydrogen is the most common element in the universe
- No meltdown risk — the reaction stops instantly if conditions aren't maintained, rather than running away
- Minimal long-lived radioactive waste — nothing like the waste storage problem fission plants face
- No carbon emissions during operation
The catch is that fusion has never been commercially viable. Getting hydrogen atoms to fuse requires temperatures hotter than the sun's core, and for decades the energy needed to trigger the reaction exceeded the energy it produced. That's why fusion has a well-earned reputation as the technology that's "always 30 years away."
Proxima is pursuing a specific approach called a stellarator — a twisted, pretzel-shaped magnetic chamber that confines superheated plasma. It's mechanically more complex to build than the donut-shaped "tokamak" design other fusion companies use, but stellarators are theoretically more stable once running, which could matter a lot for a power plant meant to operate continuously for years.
Google Isn't Fusion's Only Big Tech Backer
This is where the story gets bigger than one funding round. Google's Proxima investment is its second major fusion bet — it's also backed Commonwealth Fusion Systems (CFS), a Massachusetts-based company building a tokamak reactor, and signed a deal in 2025 to purchase 200 megawatts of power from CFS's planned Virginia plant once it's operational.
Google isn't alone:
- Microsoft signed a power purchase agreement with Helion Energy, the fusion startup backed personally by OpenAI CEO Sam Altman, which claims it will start supplying Microsoft with electricity by 2028
- Amazon has taken a different nuclear path, investing in X-energy to build small modular fission reactors — not fusion, but part of the same broader push to lock in dedicated power supply
- OpenAI itself has warned that "electrons are the new oil," and has reportedly told U.S. officials it's worried about an "electron gap" limiting how fast it can grow
None of these fusion plants exist yet. The earliest demonstrators are targeting the early 2030s, and full commercial plants are pointed at "later this decade" at best — which in an industry with a long history of delays, is a hopeful way of saying "we're not entirely sure."
Why Now? The AI Power Problem, Explained
This surge in fusion investment isn't happening in a vacuum. It's a direct response to how much electricity AI now requires.
Training and running large AI models demands enormous, constant power — data centers full of GPUs running around the clock, plus the cooling systems needed to keep them from overheating. Google's own environmental disclosures showed data center electricity use jumping sharply this year, and the pattern is the same across Microsoft, Amazon, and Meta.
The problem is that the U.S. power grid wasn't built for this. It's an aging, loosely connected patchwork, and adding new transmission capacity can take the better part of a decade. Renewables can be deployed faster than new gas plants, but they don't provide the constant, "always-on" baseload power that a data center running AI workloads needs.
Faced with that bottleneck, tech companies are throwing capital at every possible fix at once: expanding renewables, adding battery storage, restarting old nuclear fission plants, building small modular reactors — and making long-shot bets on fusion, in case it finally works this time.
The Skeptics Have a Point, Too
It's worth being honest about where things stand. Fusion has produced genuine scientific milestones — the U.S. National Ignition Facility achieved a net energy gain in a lab setting back in 2022 — but going from a lab experiment to a power plant that reliably feeds a electrical grid is a completely different challenge.
Unresolved problems include:
- Whether reactor materials can survive years of neutron bombardment without degrading
- Securing a steady supply of tritium, a rare hydrogen isotope fusion reactors need as fuel
- Whether fusion electricity can ever be cheap enough to compete with increasingly affordable solar, wind, and battery storage
Google itself has been candid about this. In discussing its CFS partnership, the company said commercializing fusion is "immensely challenging, and success is not guaranteed" — not exactly the language of a sure bet.
What This Means for the Rest of Us
Even if you'll never see a fusion plant with your own eyes, this race matters in a few practical ways:
- Your electricity bill — data center demand is already pushing up regional power prices in some markets, and how quickly new supply comes online affects that
- Where AI infrastructure gets built — tech companies increasingly site data centers based on power availability, not just internet connectivity or tax breaks
- Climate impact — if fusion works at scale, it could meaningfully change the emissions math on the compute-heavy AI industry; if it doesn't, expect more gas and fission in the near term instead
For now, the honest takeaway is this: Big Tech is making venture-scale bets on a technology that has repeatedly disappointed its believers, because the alternative — running out of power to train the next generation of AI models — is a risk they're apparently even less willing to take.
Frequently Asked Questions
Is nuclear fusion the same as nuclear power plants we have today?
No. Today's nuclear plants use fission, which splits atoms apart. Fusion combines atoms together, the same reaction that powers stars. Fusion doesn't carry meltdown risk and produces far less long-lived radioactive waste.
When will fusion power actually be available?
Not soon. Proxima Fusion is targeting an early-2030s demonstrator, with a commercial plant "later this decade" at the earliest. Other companies, like Commonwealth Fusion Systems and Helion, have similar or slightly earlier timelines. No fusion power plant has ever supplied electricity to a commercial grid.
Why is Google investing in fusion instead of just building more solar or wind?
Google is doing both. Renewables are faster to deploy but can't provide constant baseload power the way a fusion or fission plant theoretically could. Fusion is a long-term, high-risk, high-reward bet layered on top of nearer-term renewable and storage investments.
How much has Google invested in fusion overall?
Google hasn't disclosed a combined total. It contributed to Proxima Fusion's $468 million round in July 2026 and has an undisclosed equity stake in Commonwealth Fusion Systems, alongside a 200-megawatt power purchase agreement tied to CFS's planned Virginia plant.
This article is based on confirmed statements from Google, Proxima Fusion, and reporting from CNBC, Bloomberg, and Reuters as of July 8, 2026. Fusion plant timelines are company projections, not guarantees, and have historically been subject to delay across the industry.
