Crusoe, the Denver-based AI data center builder, has ended its $1.25 billion deal to buy natural gas turbines from Boom Supersonic. TechCrunch AI reports that Boom CEO Blake Scholl confirmed the split Friday in a post on X. Crusoe also confirmed to TechCrunch that it’s no longer doing business with Boom.
This matters for two reasons. First, it shows how fast power plans change at AI data centers. Second, Boom has lost the launch customer for a business meant to pay for its supersonic jet.
📋 The deal
The original agreement included:
- Buyer: Crusoe, which started in 2018 as a bitcoin miner running on excess natural gas from oil fields. It recently raised $3.9 billion.
- Seller: Boom Supersonic, which is developing the Overture supersonic passenger jet.
- Hardware: 29 of Boom’s “Superpower” turbines, each rated at 42 megawatts. That’s about 1.2 gigawatts in total.
- Price: $1.25 billion.
- Timeline: Deliveries were set to start in 2027.
The Superpower turbine shares about 80% of its parts with Symphony, the engine Boom is building for Overture. The idea was to take jet engine technology, run it on natural gas on the ground, and sell it to power-hungry data centers.
🎯 Why Crusoe walked
Scholl kept his explanation short.
“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.”
Crusoe’s reply sounded diplomatic but noncommittal. Spokesperson Andrew Schmitt said the company stays flexible and picks “the energy solutions that are right for each site as its needs evolve,” including turbines, wind, solar, batteries and the grid. “While Boom has been a great partner, the partnership isn’t the right fit today,” he added.
Crusoe’s own sites show how mixed the approach really is:
- Abilene, Texas (1.2 GW, built for Oracle and OpenAI): Runs on grid power. A gas turbine plant exists only as backup.
- Abilene, Texas (900 MW, built for Microsoft): Will run on on-site gas turbines.
So Crusoe still uses turbines. It just doesn’t need Boom’s turbines as a main power source right now.
⚠️ Damage assessment for Boom
Boom takes the bigger hit here. The company raised $300 million last year, mostly to get the turbine business off the ground. Scholl told TechCrunch at the time that turbine profits would help fund Overture’s development.
Losing your first and headline customer raises questions about demand. Scholl says the pipeline is still healthy:
- About 250 MW of Superpowers will go to other sites next year.
- Boom is aiming for 1 GW of deliveries in 2028.
- He left room for a future deal with Crusoe “if/when turbines become part of their primary power mix.”
Those targets are ambitious for a company whose turbine hasn’t built a delivery track record yet. Watch whether Boom names new customers publicly. Unnamed buyers in a pipeline don’t carry the same weight as a signed $1.25 billion order.
🔍 Why this matters for the AI industry
What stands out here is how unsettled power strategy still is, even for the biggest AI infrastructure builders. For the past two years, on-site gas generation has looked like the fix for a clear problem: grid connections can take years, and AI campuses need gigawatts now. That demand opened the door to newcomers like Boom that are repurposing aviation hardware.
This deal shows the other side. Operators like Crusoe are hedging across grid power, gas, renewables and batteries, site by site. A turbine order that made sense a year ago can drop out once grid power comes through or site plans shift.
For practitioners and buyers of AI compute, a few points to take away:
- Power is still the bottleneck. Data center capacity depends on energy deals that can change quickly.
- Big announced deals aren’t final. A $1.25 billion commitment disappeared before a single turbine shipped.
- New energy suppliers carry execution risk. Companies entering the AI power market from other industries depend heavily on a few anchor customers.
🧭 Outlook
Crusoe keeps growing with fresh capital and a flexible energy strategy. Boom now has to prove its turbine business can stand without its launch partner, and its 2027 and 2028 delivery targets are the next checkpoints. Full details are available in the original TechCrunch AI report.