AI data centers are turning to converted jet engines for electricity generation, bypassing traditional grid connections that face multi-year delays. The strategy allows facilities to become operational within 24 months—critical in a race for computing capacity.
The speed differential is stark. Connecting 100 megawatts to the grid in Northern Virginia takes an estimated seven years, according to the International Energy Agency. Most new power capacity nationwide is not expected online until 2031, an executive at Mitsubishi Power confirmed.
These delays are already forcing operational decisions. Microsoft withdrew as the anchor customer for Nscale's data center in Loughton, England, after the local utility said it could not deliver sufficient power until the early to mid-2030s.
xAI has already deployed this strategy at scale. The company installed eight converted jet-engine turbines in Mississippi to power its Colossus 2 data center in Memphis, according to state permits and satellite imagery.
OpenAI followed suit in 2025, deploying gas turbines at its Stargate site in Abilene, Texas. Crusoe, a Denver-based data center developer building the Abilene campus, ordered 13 converted units from ProEnergy in April.
ProEnergy rebuilds General Electric engines salvaged from Boeing jets for power generation. The rapid deployment timeline—months rather than years—makes these units attractive to operators under time pressure.
Crusoe had pursued a larger acquisition: a $1.25 billion deal with Boom Supersonic for 29 natural gas turbines. The company terminated that agreement in September, though Crusoe said it would continue using turbines from other suppliers.
The Trump administration has encouraged data centers to generate their own electricity, adding policy tailwinds to the self-sufficiency push. But the economics are a trade-off. Converted jet engines are less efficient than turbines designed for stationary power plants and produce higher emissions. Operators need more units to generate equivalent output.
Industry analysts expect these efficiency and environmental concerns to intensify within the next one to two years as the scale of deployment grows. For now, data center operators are choosing speed and capital deployment over long-term cost optimization and emissions footprint—a calculation likely to shift once grid bottlenecks ease.
