Global investors poured a record $4.48 billion into fusion energy companies over the past year, and four startups absorbed more than half of it.
Commonwealth Fusion Systems, Inertia Enterprises, Helion Energy and Proxima Fusion between them took over half of the total, according to a survey from the Fusion Industry Association reported by Reuters and Bloomberg. The number is up 69% from a year earlier. Four companies. More than half the market. That kind of concentration tells you something about where investors think the real bets are.
Commonwealth Fusion Systems alone closed a $3.85 billion round in May, pushing its lifetime funding to roughly $6.85 billion since it spun out of MIT in 2018. Its SPARC tokamak, under construction in Devens, Massachusetts, is about 75% complete by the company's own account. CFS already has skin in the AI power game too. It signed a 200 megawatt power purchase agreement with Google last year, and Eni followed with a deal worth more than $1 billion in September 2025.
Helion Energy raised $465 million in a June Series G that valued it at $15.5 billion, bringing total funding to $1.5 billion. Sam Altman put $375 million of his own money into Helion back in 2021, still his largest personal investment. Helion is contracted to sell Microsoft 50 megawatts of fusion electricity by 2028, enough to power roughly 40,000 homes in Washington state if the machine hits its numbers. Nobody has built a commercial fusion plant yet. That contract is a bet, not a delivery.
Proxima Fusion, based in Munich, took a different technical path. Rather than the tokamak most rivals chase, it is building a stellarator, a twisted reactor shape designed to hold plasma stable without constant magnetic correction. Google's growth investing arm backed a $468 million round for the company this month, according to CNBC. Inertia Enterprises is the newest of the four. It came out of stealth in February with $450 million in Series A funding led by Bessemer Venture Partners, joined by GV, Modern Capital and Threshold Ventures. Its founders are veterans of the National Ignition Facility, the government lab that achieved fusion ignition with lasers in 2022, and the company is chasing that same laser-driven approach commercially.
The Money Isn't Reaching Suppliers #
Fusion developers have now raised $14.2 billion since the Fusion Industry Association started tracking the number in 2021. This single year accounts for close to a third of everything the industry has ever raised. But the money is not spreading evenly. A separate FIA supply chain survey of 25 fusion companies and 67 suppliers found something less encouraging: 69% of suppliers say a lack of large, long-term orders from developers is the main thing holding back further investment in fusion hardware. Supply chain spending still rose 24%, to $538 million. The suppliers who make the superconducting wire and the cryogenic systems are stuck waiting on the same four or five companies to place orders big enough to justify building capacity. The money moved. The orders didn't. That's the chicken and egg problem nobody in fusion has solved yet.
Fusion Goes Public #
General Fusion gave the sector its first real market test of investor appetite outside venture capital. The Vancouver company is backed by Jeff Bezos and Shopify co-founder Tobias Lütke. It completed a reverse merger with Spring Valley Acquisition Corp. III around July 10, becoming the first pure-play fusion company to trade on Nasdaq, under the ticker GFUZ. That's a first. The deal delivered $105 million in public equity investment plus $230 million from Spring Valley's trust, valuing the combined company near $1 billion. Public markets can now put a daily price on fusion risk, not just venture partners voting with term sheets once a year.
The AI power story is doing real work here, not just providing atmosphere for the pitch decks. Data center electricity demand in the US could reach 35 to 45 gigawatts by 2030, roughly double 2024 levels. Hyperscalers have already committed to buying fusion power years before anyone has proven a commercial plant works. Google and Eni did not sign those deals with CFS out of charity. They did it because the alternative, building enough new gas and nuclear capacity fast enough to feed AI clusters, looks just as hard. Whether that translates into actual electrons on the grid by 2028 is still an open question. The industry's own suppliers are the ones saying so loudest.
Also read: Delaware Weighs a New Legal Entity Built for Companies Run by AI Agents • Google's Money Just Made Proxima Fusion Europe's Best Funded Fusion Bet • Klarna Wants Its Own Bank, and Regulators Now Have to Decide