Stuck in years-long lines for grid connections and new turbines, some AI data centers are making their own electricity with converted jet engines, the New York Times reported Tuesday.
The engines, pulled from retired aircraft and rebuilt to burn natural gas, aren't anyone's idea of a permanent power plant. But they're another sign of how far power supply has fallen behind the AI buildout.
Why Data Centers Are Looking at Jet Engines
Companies that need massive, reliable electricity for AI workloads are hitting a bottleneck. The International Energy Agency says hooking up 100 megawatts to the grid in Northern Virginia takes about seven years. By contrast, a converted jet-engine unit can be operational in 24 months, according to Rystad Energy analyst Audun Martinsen.
"It's all about getting power as fast as possible," Martinsen told the Times. The speed of deployment matters because AI projects often have multi-year timelines and investors expect rapid scaling.
How the Conversions Work
Most of the engines being repurposed are GE models that once powered Boeing jets. Companies such as ProEnergy dismantle the aircraft, refurbish the core turbine, and replace the original aviation fuel system with a natural-gas burner. The resulting unit resembles a small gas turbine plant but retains the high-speed compressor and turbine stages designed for jet thrust.
Because the hardware was engineered for high power-to-weight ratios, the converted units can produce roughly 10-12 megawatts each, depending on configuration. To reach a 100-megawatt capacity, a data center would need about eight to ten of these engines, whereas a conventional combined-cycle gas plant might need only two larger units.
Key Deployments So Far
Elon Musk's xAI has installed eight converted jet-engine turbines in Mississippi to power its Colossus 2 data center in Memphis. The installation is documented through state permits and satellite imagery, the Times reported.
OpenAI plans to use similar gas turbines at its first Stargate campus in Abilene, Texas, slated for 2025. The Denver-based startup Crusade (styled as Crusoe) is building the Abilene campus and ordered 13 ProEnergy-converted units in April.
In September, Crusoe walked away from a $1.25 billion contract to buy 29 natural-gas turbines from Boom Supersonic. Boom, known for developing a supersonic passenger jet, also markets a version of its jet engine as a power-plant for data centers. A Crusoe spokesperson told TechCrunch that the company still intends to use turbines, though the source may change.
Economic and Regulatory Context
The United States is seeing a surge in AI-driven data-center construction. At the same time, the power sector struggles to keep pace. An executive at Mitsubishi Power, one of the world's leading makers of power-generation equipment, told Bloomberg that most new capacity is expected to come online by 2031.
Regulatory pressure is also shifting. The current administration is encouraging data centers to generate their own electricity, a policy stance that reduces reliance on strained utility networks but raises questions about emissions and local permitting.
Environmental and Efficiency Trade-offs
Converted jet engines are less efficient than purpose-built gas turbines. The Times notes they pollute more and require a larger number of units to deliver the same output. Experts warn that as the AI market matures, efficiency and environmental impact will become more significant factors within a year or two.
Because the engines burn natural gas, they emit carbon dioxide and nitrogen oxides at higher rates per megawatt than modern combined-cycle plants. Communities near installations may see increased local air-quality concerns, prompting tighter permitting reviews in some states.
Impact on Data-Center Operators and the Broader Market
Delays in power delivery are already affecting business decisions. Last month, Microsoft withdrew as the anchor tenant for Nscale's data-center project in Loughton, England, after the local utility indicated it could not provide sufficient power until the early-to-mid 2030s.
Operators that adopt jet-engine solutions can avoid such setbacks, but they must balance short-term speed against longer-term cost and compliance risks. The capital cost of a converted engine is lower than a full-scale gas plant, yet the higher fuel consumption may raise operating expenses.
What Comes Next for the Industry
Analysts expect the jet-engine approach to remain a stop-gap measure while grid upgrades and new turbine projects progress. The International Energy Agency's timeline suggests that most of the required capacity will be in place by the early 2030s.
Stakeholders should monitor several indicators: the issuance of new permits for jet-engine conversions, changes in state-level emissions regulations, and the pace of traditional gas-turbine deliveries from manufacturers like Mitsubishi Power and GE.
In the meantime, data-center developers are likely to continue evaluating a mix of power-source options, including renewable-energy contracts, battery storage, and on-site generation, to meet the urgent demand for AI compute power.
