Some AI data centers are turning retired aircraft engines into power sources as they face years-long waits for grid connections and new turbines. Rebuilt to burn natural gas, the engines offer a way to get electricity sooner, but the trade-off is lower efficiency and more pollution.
Speed matters when grid power is delayed
For operators trying to build large data centers, the attraction is time. Rystad Energy analyst Audun Martinsen told the New York Times, “It’s all about getting power as fast as possible.” He said jet-powered units can have a data center running in 24 months.
Grid access can take much longer. In Northern Virginia, connecting 100 megawatts to the grid takes about seven years, according to the International Energy Agency. That gap helps explain why some companies are willing to use equipment that was originally built for aircraft rather than conventional power plants.
The delays have commercial consequences, too. Last month, Microsoft pulled out as the anchor customer for Nscale’s data center in Loughton, England, after the local utility said it could not provide enough power until the early-to-mid 2030s. A project without the electricity it needs may struggle to move forward on the schedule its customers expect.
Companies are building power on site
xAI is among the companies already using converted jet engines. The New York Times reported, citing state permits and satellite imagery, that the company installed eight converted jet-engine turbines in Mississippi to help power its Colossus 2 data center in Memphis.
The units come from ProEnergy, which rebuilds a GE engine commonly found on Boeing jets. The conversion turns an engine from retired aircraft into equipment that can generate electricity by burning natural gas.
OpenAI also moved toward gas turbines in 2025, following Musk’s lead at its first Stargate site in Abilene, Texas. Crusoe, the Denver-based startup building AI data centers, is constructing that campus and ordered 13 of ProEnergy’s converted units in April.
Crusoe’s plans have shifted as it weighs different turbine options. In September, it walked away from a $1.25 billion deal to buy 29 natural gas turbines from Boom Supersonic. The startup, which is developing a supersonic passenger jet, has also begun selling a version of that jet’s engine as a power plant for data centers. A Crusoe spokesperson told TechCrunch at the time that the company still planned to use turbines.
A fast option with efficiency costs
Converted aircraft engines are not designed to be permanent power plants. According to the New York Times, they are less efficient and pollute more than turbines built for conventional gas plants. Operators also need more of them to produce the same amount of power.
Those limits make the choice a practical compromise. A data center operator may value the ability to bring a site online sooner, especially when waiting for grid service could delay a project or lose a customer. But generating power on site does not remove the costs of using equipment that produces less electricity efficiently and creates more pollution.
The Trump administration is adding to the pressure by encouraging data centers to generate their own electricity. At the same time, experts say efficiency and environmental concerns could become more important in a year or two. If that happens, the speed that makes converted engines appealing today may have to be weighed more heavily against their operating drawbacks.
Power supply is shaping the AI buildout
The move to rebuilt jet engines points to a broader challenge: the supply of power has not kept pace with the expansion of AI data centers. The equipment choice reflects the urgency of getting electricity to new sites while grid connections and new turbine capacity remain difficult to secure quickly.
For now, converted engines offer one route to faster power. Their use also shows that securing electricity is no longer just a question of connecting a data center to the grid; in some cases, companies are bringing generation equipment to the site and accepting the efficiency and pollution trade-offs that come with it.