xAI, Crusoe tap retired jet engines to power AI data centers

Elon Musk's xAI has installed eight converted jet turbines to power a data center in Mississippi, according to state regulatory filings and satellite imagery reviewed by The New York Times.

The driver is a wait. New gas turbines take six years on average to deliver in the United States, and connecting even 100 megawatts to the grid in Northern Virginia takes about seven years, according to the International Energy Agency. So builders are going to the boneyards.

ProEnergy takes the CF6-80C2 engine, the kind that powers a Boeing 767, and turns it into a power turbine of up to 48 megawatts. Crusoe, the data center builder constructing the Stargate campus in Texas for OpenAI, ordered 13 of them in April, for a combined 650 megawatts.

"It's all about getting power as fast as possible," Rystad Energy analyst Audun Martinsen told the paper. With jet-powered units, "they can have a data center running in 24 months."

The raw inventory is enormous. The roughly 4,000 aircraft stored at the Arizona boneyard could, in theory, represent as much as 40 gigawatts of capacity, though that figure assumes engines remain intact and serviceable after an average of more than a decade in storage. Three of the analysts on our tracker had posted the piece by the same evening.

Not everyone sees progress. Patrick Anderson, an attorney with the Southern Environmental Law Center, told the paper that electric generation for AI "should be getting cleaner over time, not dirtier." Converted jet engines are generally less efficient and far more polluting than modern turbines.