Crusoe Cancels $1.25 Billion Gas Turbine Order With Boom Supersonic
Crusoe Energy has unwound a $1.25 billion agreement to buy 29 natural-gas turbines from Boom Supersonic's Superpower division, ending a launch-customer relationship the two companies announced only nine months earlier. The split was confirmed independently by TechCrunch, Data Center Dynamics and Boom's own public statements, and it marks one of the largest reversals yet in the scramble by AI data center developers to lock down behind-the-meter power generation.
The original order, struck in December 2025, called for 29 of Boom's 42-megawatt Superpower units — a combined 1.21 gigawatts — with first deliveries targeted for 2027 at Crusoe's Abilene, Texas campus, the site built for OpenAI and Oracle's Stargate project. The deal coincided with Boom's $300 million Series B, led by Darsana Capital Partners, which the company said would fund commercialization of turbines derived from the same high-temperature materials as its Symphony jet engine.
Crusoe's reversal is notable for its timing rather than any sign of financial distress: the company closed a $3.9 billion Series F on September 17, 2026, reaching a $30.9 billion post-money valuation, just eight days before confirming the turbine deal was off. According to Crusoe spokesperson Andrew Schmitt, the Abilene campus built for OpenAI and Oracle is now primarily grid-powered, with gas turbines relegated to backup, while a separate 900-megawatt Crusoe site for Microsoft remains planned around on-site gas generation — from suppliers other than Boom.
"The TL/DR is that turbines are no longer part of Crusoe's near term primary power mix at Abilene/etc., so a launch partnership just didn't make sense." — Blake Scholl, Founder and CEO, Boom Supersonic
The cancellation lands amid an industry-wide crunch in exactly the equipment Crusoe walked away from: turbine manufacturers including GE Vernova have reported order backlogs stretching toward 2030 as data center operators race to secure generation capacity ahead of grid interconnection queues that can run years long. That Crusoe chose to exit a committed position with a pre-commercial, aviation-derived entrant — rather than hold the slot as insurance — suggests operators are increasingly weighing supplier track record and delivery certainty against simply reserving scarce capacity early.
For the broader market, the split reads less as a retreat from on-site power and more as evidence that procurement has matured into a portfolio exercise. Crusoe's own language — describing a mix of turbines, wind, solar, batteries and grid supply evaluated site by site — mirrors how other large developers are now treating power sourcing as a live, renegotiable variable rather than a single long-term bet. Boom, for its part, says it still holds roughly $1 billion in backlog and expects to deliver 250 megawatts to other customers in 2027, indicating that demand for novel on-site generation has not disappeared, even if Crusoe decided the risk profile no longer matched its needs.