xAI's eight converted jet engines show how far AI power has fallen behind

The queue behind the jet engines

The units come from ProEnergy. The company rebuilds a GE engine commonly found on Boeing jets, according to the New York Times reporting that TechCrunch cited on 3 November. State permits and satellite imagery identified the installation in Mississippi.

Power is now the binding constraint on new AI capacity. Hooking up 100 megawatts to the grid in Northern Virginia takes about seven years, according to the International Energy Agency. Neither report gives a figure for the power the eight units deliver, or for how long they will run.

Rystad Energy analyst Audun Martinsen told the Times that speed is the whole point. "It's all about getting power as fast as possible," he said, adding that with jet-powered units "they can have a data center running in 24 months". Most power capacity is expected to come on by 2031, an executive at Mitsubishi Power told Bloomberg last week, as TechCrunch reported. The delays are already costing operators business. Last month, Microsoft pulled out as anchor customer for Nscale's data centre in Loughton, England, after the local utility said it could not deliver enough power until the early to mid-2030s.

Who else is buying turbines

OpenAI used gas turbines at its first Stargate site in Abilene, Texas, in 2025. Crusoe builds that campus. The Denver startup ordered 13 of ProEnergy's converted units in April.

In September, Crusoe walked away from a $1.25 billion deal to buy 29 natural gas turbines from Boom Supersonic. Boom is developing a supersonic passenger jet and has begun selling a version of that jet's engine as a power plant for data centres. A Crusoe spokesperson told TechCrunch at the time that the company still plans to use turbines.

The pattern extends beyond aircraft engines. TerraVolt acquired a site in Idaho to develop a 240 MW gas-powered behind-the-meter campus, DataCenterDynamics reported on 3 November. California Resources and Bloom Energy are eyeing 1 GW of fuel cells across California for data centres and other offtakers, the same outlet reported.

The pollution bill

Permits for the Mississippi turbines allow a 25% higher limit for volatile organic compounds. The Southern Environmental Law Center has challenged the state's decision. It has sued xAI before over gas generators.

The group argues that xAI has not guaranteed the turbines meet emissions standards for pollutants including formaldehyde, and accuses the company of undercounting projected emissions relative to the data centre's power sources. Converted jet engines are less efficient and pollute more than turbines designed for conventional gas plants, according to the Times. It also takes more of them to make the same amount of power. Some experts expect efficiency and environmental concerns to grow in importance in a year or two.

Washington pushes onsite, equipment runs short

In an Oct. 21 Truth Social post, President Donald Trump said new, large-scale data centres are "mandated by me to make their own electricity". He said they should sell excess power back to the grid.

Equipment is the obvious obstacle. Many components for gas-fired, nuclear and battery-backed onsite systems are in tight supply, with some turbine models effectively sold out through 2031, Data Center Knowledge reported. Patrick Hughes, senior vice president at the National Electrical Manufacturers Association, said a standalone AI data centre can technically power itself. But that does not make it feasible for a single site, let alone the whole fleet, to run on its own power around the clock, he said. Hughes said the most likely outcome for many facilities is a hybrid approach that combines onsite generation and storage with a grid connection.

The money behind the rush

The turbine trade has a cost for the companies selling it. Ars Technica reported on 1 November that United Airlines' chief executive doubts Boom's supersonic airliner will be ready by 2029. Boom has been shifting toward powering AI data centres.