Product1 distinct publisher3 min readPublished
The Wisconsin utility put money into Realta Fusion and offered it somewhere to plug in, for a plant that would not run until the mid-2030s. That is a wager on how short of power the next decade gets.
The Product Desk · Product desk
Compiled by The Product DeskSomething wrong?How this is made
Somebody at Madison Gas and Electric keeps the long-range resource plan, and the mid-2030s page of it now carries a 200 megawatt line attached to a plant that does not exist [2]. What does exist is a research and development shop in a converted Oscar Mayer factory in Madison [14].
Separate what the utility bought from what fusion is selling. MGE bought a shareholder's position in Realta and an early look at a technology it might want in a decade [3]. It paid with the things a utility has and a startup cannot buy: somewhere to interconnect, plus engineering hours and financing for the plant itself [4]. In the two deals furthest along, the utility plays a narrower part. Dominion leases land to Commonwealth Fusion Systems for Arc, a 400 megawatt plant near Richmond expected online in the early 2030s, and Google and the Italian energy company Eni signed to buy the output [7][8]. Helion leases from the Chelan County Public Utility District in Washington and sells to Microsoft [9]. In both, someone other than the utility carries the price risk. TechCrunch's account of the Wisconsin arrangement names no buyer for the electricity [18], and among the American deals it describes, MGE is the only utility reported to have put in equity [17].
Add up the numbers and the picture sharpens. Realta's 200, Arc's 400, Type One's 350 and Helion's 50 come to 1,000 megawatts of announced US fusion capacity [15]. Only Helion's 50 is meant to run before 2030, which leaves 950 megawatts, 95% of the total, landing in the 2030s [16]. None of this cohort is selling into today's shortage. They are selling into the one utilities say they see coming, as AI data centre demand meets a hunt for round-the-clock supply that is not fossil-fuelled [13][12].
It helps to sort these deals along two axes: whether you hand the counterparty a scarce physical asset such as land or a queue position, and whether anyone commits to a price for the output. Dominion and Chelan sit in the box marked asset given, price set by a third party [7][9]. MGE sits in a different box: asset given, output unpriced for now, which is the cheapest box to enter and the hardest to value afterwards.
The forcing function is the counterfactual on the site: what that interconnection position would have carried had Realta never built, and in what year it would have carried it. If the honest answer is nothing before 2035, the option cost about what the equity cost, and this is a sensible cheap bet. If the answer is a battery or a peaker that wanted the slot in 2030, MGE has paid a price that appears in no line item, and the same test applies to any partnership you are being offered on a technology that works on paper.
Ranked by verification strength, evidence, and original report placement.
Realta Fusion announced a deal with Madison Gas and Electric that will place one of the first grid-connected fusion power plants in the country in Wisconsin.
The Realta deal with Madison Gas and Electric will explore the construction of a 200 megawatt power plant, roughly enough capacity to power a small city, sometime in the mid-2030s.
Madison Gas and Electric also made an equity investment as part of the deal; Realta did not reveal specifics and did not immediately reply to TechCrunch's request for more information.
Realta gains access to interconnection sites, engineering and technical assistance, and financing for the eventual power plant.
Access to interconnection sites is a key advantage because power providers and users are scouring the country for a place to hook up to an increasingly connected electrical grid.
To date only a handful of fusion startups have announced deals with utilities, which is what makes agreements like this one newsworthy.
Distinct publishers with included, body-backed reporting in this cluster.
1 article · September 3, 2026
Follow any of these and your For You feed starts watching them — no settings page required.
invest
Corporate venture arms are writing deep tech's first purchase order1 distinct publisher
build
Grok 4.6 lands in Copilot two days after launch, and the model picker becomes a procurement problem1 distinct publisher
product
Washington's secret AI test is coming for open weights, and release dates go with it2 distinct publishers
invest
Nvidia's August 26 print: 92% of the quarter rides on one segment1 distinct publisher
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
One outlet, one announcement
Everything about Wisconsin rests on TechCrunch's account of a company announcement, and the detail that would matter most — the size of Madison Gas and Electric's stake — is the one Realta declined to give, as TechCrunch says plainly. The four comparison projects are reported crisply and are checkable against earlier public announcements, which keeps the floor off the ground. The deal at the centre has no second telling, no filing, and no quoted word from the utility.
Announcements, not electrons
No fusion plant sells power to any grid today, and this reporting does not pretend otherwise. What exists is five announced pairings and a meat-plant conversion in Madison. On the dates given, 950 of the 1,000 announced US megawatts land in the 2030s; Helion's 50-megawatt Polaris, targeted for 2028, is the only near-term test of whether any of it becomes electricity.
The tense is doing the work
'One of the first grid-connected fusion power plants in the country' describes a plant being explored, not built, for a decade out. TechCrunch is careful with 'explore' and with the mid-2030s date, so the overstatement is structural rather than sloppy: a race framing rewards announcement velocity, and an option written up beside signed offtakes starts to read like a commitment. A 200-megawatt plant with no buyer named and an equity check of unknown size is a wager, and it is presented as a milestone.
Everyone gains from the headline
The announcement is the source. Realta converts a utility's name into fundraising credibility — TechCrunch says as much, calling these tie-ups a credibility check for founders and investors — while Madison Gas and Electric buys cheap optionality and a clean-supply narrative. The double role recurs abroad, where RWE is simultaneously Proxima's landlord and its investor. Nobody in the story has reason to volunteer doubt, and the question that would test the enthusiasm went unanswered.
Direction solid, numbers thin
That cautious utilities are now at the table is easy to believe and reasonably well observed. Beyond that there is little to hold: one outlet, an undisclosed stake, a target date in the mid-2030s, and no description of the technology being bet on. A stated investment amount, a rate filing, or a word from Madison Gas and Electric itself would move this materially.