AI Data Centers Are Turning Retired Jet Engines Into Power Plants
Some AI data centers now generate their own electricity with gas turbines built from retired jet engines, because developers cannot wait for grid connections and turbine deliveries that can take years. TechCrunch reported the trend on November 3, 2026.
The units start as aircraft engines. A gas turbine is a machine that uses burning fuel to spin a shaft that drives a generator, and the converted units are rebuilt to burn natural gas. ProEnergy supplies converted turbines built by rebuilding a GE engine commonly found on Boeing jets. FTAI Aviation repurposes CFM56 engines into 25 MW (megawatt) turbines for AI and data center power. FTAI President David Moreno has said the conversion of a jet engine into a power-generating unit takes the company 30 to 45 days.
The customers include xAI, which installed eight converted jet-engine turbines in Mississippi to help power its Colossus 2 data center in Memphis. Crusoe, a Denver-based AI data center startup that had recently raised $3.9 billion as of September 2026, ordered 13 of ProEnergy's converted generating units in April for its AI factories, including the Abilene, Texas campus.
The draw is schedule. Rystad Energy analyst Audun Martinsen said speed is the main attraction of converted jet engines, and that jet-powered units can have a data center running within 24 months. The grid route is much slower. According to the International Energy Agency, hooking up 100 megawatts to the grid in Northern Virginia takes about seven years.
Grid constraints are forcing other changes in project plans. Microsoft pulled out as anchor customer for Nscale's data center in Loughton, England, after the local utility said it could not deliver enough power until the early-to-mid 2030s. In September 2026, Crusoe walked away from a $1.25 billion deal to buy 29 natural gas turbines from Boom Supersonic.
The trade-off is in the hardware. Converted jet engines are less efficient than turbines designed for conventional gas plants and produce more pollution per unit of output, so producing the same amount of power requires more of them. Aeroderivative turbines, which are derived from aircraft engine designs, offer rapid start-up times and high power density, meaning a lot of output from a compact machine. That is why they suit bridging utility delays. GE Vernova offers onsite generation for data centers, spanning simple-cycle aeroderivative units to heavy-duty combined-cycle systems.
The wider market is moving in the same direction. Siemens Energy reported record third-quarter sales, margins and orders in August 2026, driven by demand for gas turbines from expanding AI data centers. Reuters has reported that the dash for small gas turbines is set to raise long-term power costs and emissions for data centers. The International Energy Agency projects that global data center electricity use will roughly double by 2030. Batteries, meanwhile, are now cheaper than the natural gas turbines used at many data centers, according to TechCrunch's October 2026 coverage.
The broader context here is one of sequencing. A bridge generator that comes online in two years can carry a campus through the period before a grid connection arrives. For an operator facing a seven-year queue, that difference can decide the entire project economics. The emissions and heat-rate penalties, however, run for as long as the machines do. Whether these units are retired once grid capacity arrives, or quietly become baseload power, will determine how much of the speed advantage is temporary.
In my view, the more useful question for engineers is not whether converted engines work, since they clearly do in service, but what gets built around them. Onsite generation changes siting decisions, fuel logistics and permitting. It also moves the bottleneck from the utility's substation to the operator's own gas supply and maintenance schedule. If those constraints are managed well, the approach can bring compute online years sooner than the grid would allow, which is a real gain for model training and inference capacity.
