Crusoe Just Killed Its $1.25B Boom Turbine Deal — and Exposed the Real Bottleneck in the AI Buildout
Nine months ago, it was the deal that proved AI data centers didn't have to wait for the grid. Crusoe, the Denver AI infrastructure company racing to build gigawatt-scale computing campuses, signed on as the launch customer for Boom Supersonic's "Superpower" natural-gas turbines — 29 of them, 1.21 gigawatts of generation, a package valued at more than $1.25 billion. First deliveries were due in 2027.
That plan is now dead.
On Friday, September 25, Boom Supersonic founder and CEO Blake Scholl announced on X that the two Denver companies are "no longer moving forward with the turbine launch partnership." Crusoe confirmed to TechCrunch that the companies are no longer doing business under the agreement. The breakup was cordial, almost friendly — and it says something important about where the real constraints in the AI boom actually sit.
A partnership unwound in a single post
Scholl broke the news himself, and he did it with unusual grace. The post opened by congratulating Crusoe founders Cully Cavness and Chase Lochmiller on the company's recent $3.9 billion Series F raise — then delivered the bad news. The turbines, he wrote, "are no longer part of Crusoe's near term primary power mix at Abilene/etc., so a launch partnership just didn't make sense."
He followed with the numbers Boom is still promising the market: "Boom will be delivering about 250MW of Superpowers next year to other sites, and we're targeting 1GW in 2028." And he closed with the door left open: "We're grateful for the help Crusoe gave us in shaping Superpower and continue cheering for their successes. The future is long, and we look forward to potentially teaming up if/when turbines become part of their primary power mix."
No turbines ever changed hands. Crusoe never took delivery of a single unit — the first shipments were still over a year out. What died on Friday was a purchase order, not a product. But it was a purchase order that carried the weight of two companies' strategies, and losing it hurts Boom far more than it hurts Crusoe.
What the December deal was supposed to prove
To understand the significance, rewind to December 9, 2025. Boom announced that day that it had secured a backlog of more than $1.25 billion for Superpower and revealed Crusoe as the turbine's launch customer. Crusoe had ordered 29 Superpower units — 1.21 gigawatts of capacity — intended to power its AI data centers. The announcement came bundled with a $300 million Series B funding round for Boom, led by Darsana Capital Partners with participation from Altimeter Capital, ARK Invest, Bessemer Venture Partners, Robinhood Ventures, and Y Combinator.
The message was deliberate and twofold. For Crusoe, it was about escaping the utility queue: jet-derived turbines that could be trucked in and fired up in months, instead of waiting the better part of a decade for new grid interconnection. For Boom, it was a pivot story — the supersonic-jet company, best known for its Overture airliner and the XB-1 demonstrator that broke the sound barrier in early 2025, was suddenly an energy company too. Superpower revenue, Boom said at the time, would help finance the certification and delivery of Overture itself.
"Supersonic technology is an accelerant — of course for faster flight, but now for artificial intelligence as well," Scholl said in the December press release. "With this financing and our first order for Superpower, Boom is funded to deliver both our engine and our airliner." Crusoe co-founder Chase Lochmiller said the company was proud to be Superpower's launch customer, calling the initiative aligned with Crusoe's energy-first approach.
All of that architecture — the backlog, the funding narrative, the second-act story — rested on one customer. That customer is now gone.
What Superpower actually is
The hardware at the center of this story is genuinely interesting, which is part of what made the deal so compelling. Superpower is a 42-megawatt natural-gas turbine built around the engine core of Symphony, the supersonic jet engine Boom is developing for Overture. The turbine shares roughly 80 percent of its parts with the airborne engine — a clever piece of engineering arbitrage that let Boom monetize aerospace R&D in the red-hot AI infrastructure market.
According to Boom's specifications, the unit holds its full rated output in ambient temperatures exceeding 110 degrees Fahrenheit — a meaningful feature for Texas data-center sites — operates without water cooling, runs on natural gas with diesel backup capability, and arrives packaged at roughly shipping-container scale. Boom had said its total turbine production would ramp to more than 4 gigawatts annually by 2030.
None of those specifications changed between December and September. What changed was Crusoe's math on what kind of power it actually needs, and when.
Why Crusoe walked away
Crusoe's public explanation is flexibility. A company spokesperson told TechCrunch the companies are no longer doing business under the agreement, framing the shift as portfolio management: Crusoe keeps its options open on energy, choosing among turbines, wind, solar, batteries, and grid electricity as each site's requirements evolve.
That framing is worth taking seriously, because Crusoe's actual power picture at its flagship sites never matched the Superpower blueprint. The company's first 1.2-gigawatt Abilene campus, built to serve Oracle and OpenAI's Stargate buildout, draws its primary power from the grid and keeps a gas turbine plant as backup. A separate 900-megawatt Abilene campus planned for Microsoft is designed around on-site gas turbines — but Crusoe has never said those turbines were Boom's units. As TechCrunch's reporting noted, Crusoe chooses energy sources site by site, and the Superpower order was always a forward bet, not an operating commitment.
The timing also matters. Crusoe just closed a $3.9 billion Series F that reportedly valued the company at around $30.9 billion, with Atreides, Valor Equity Partners, and Mubadala co-leading the round. A company that size, building that fast, does not keep a $1.25 billion vendor commitment it no longer needs out of sentiment. Crusoe's strategy has been shifting toward factory-built, modular data centers — its Spark division is manufacturing prefabricated AI data-center modules in Colorado, Oklahoma, and Nevada — a model that favors standardized, swappable power equipment over bespoke turbine deals.
In short: the deal died because Crusoe's power plans evolved past it, not because the turbines failed. But that distinction is cold comfort for Boom.
The bigger picture: power is the real AI bottleneck
Step back from the two companies, and the breakup reads as a data point in the defining infrastructure story of the AI era. Compute gets the headlines — the GPUs, the model releases, the trillion-dollar valuations. But the binding constraint on AI expansion is increasingly electricity: who has it, where, and how fast they can get more of it.
Crusoe's pivot is a reminder that the market for that electricity is becoming ruthlessly pragmatic. Grid connections, turbines, batteries, and behind-the-meter generation are now as strategically important as chips and servers, and the developers building gigawatt campuses are treating power vendors as interchangeable parts. When site economics shift, commitments get reshuffled — quickly. A $1.25 billion headline order can evaporate into a polite X post in nine months.
That ruthlessness cuts both ways. For Boom, losing the launch customer removes the anchor order from a business that was supposed to fund the company's actual dream — the Overture supersonic airliner, whose order book stands at 130 aircraft including orders and pre-orders from United Airlines, American Airlines, and Japan Airlines. Scholl insists other customers are in the pipeline, but he hasn't named them, and his 250-megawatt-in-2027 and 1-gigawatt-in-2028 targets remain company projections, not contracted installations.
For the industry, the lesson is simpler: in the AI power race, optionality is the strategy. Crusoe is keeping every energy door open — turbines, just not Boom's — and building its campuses so the power source can change underneath them.
What happens next
Three things to watch. First, whether Boom can name a replacement launch customer before the end of the year — the credibility of its entire energy business now depends on it. Second, what Crusoe actually buys for power at its next wave of campuses, and from whom; the vendor that wins Crusoe's business after Boom will tell you who really owns the data-center turbine market. And third, whether other announced turbine-for-data-center deals — the industry has seen several in the last eighteen months — hold up once they meet the same site-by-site economics that killed this one.
The jet-engine turbine was a beautiful idea: supersonic technology as an accelerant for artificial intelligence. For now, at least, the acceleration will have to come from somewhere else.
Sources: TechCrunch (Sept. 25, 2026); TechRepublic (Sept. 28, 2026); Boom Supersonic press release via Business Wire (Dec. 9, 2025); reporting from TechInAsia, theaifeed, AI Market Watch, and Startup Fortune. Blake Scholl's quotes are from his September 25 post on X; Crusoe's statements are from its confirmation to TechCrunch. TechPulse USA previously covered Crusoe's $3.9 billion Series F and its factory-built data-center strategy.
Comments
Post a Comment