AI Data Centers Turn to Converted Jet Engines
Some operators are using rebuilt aircraft engines to generate electricity on site, bypassing long grid connection waits.
Updated on Nov. 3, 2026 in Data Centers
Live Poll
Do you expect your household electricity bills to keep rising in the next few years?
Some AI data centers are making their own electricity with converted jet engines pulled from retired aircraft and rebuilt to burn natural gas. The move comes as grid connections can take years, with hooking up 100 megawatts in Northern Virginia taking about seven years.
Why it matters
Power access has become a primary constraint on building AI data centers, and on-site generation lets operators start running facilities far sooner. The approach also raises questions about efficiency and emissions, since converted jet engines are less efficient and pollute more than turbines designed for conventional gas plants.
Converted jet-engine units can be running in 24 months, while hooking up 100 megawatts to the Northern Virginia grid takes about seven years. Crusoe ordered 13 units from ProEnergy in April, but the long-term efficiency and emissions impact remains unresolved.
The players
Microsoft
Microsoft is a major technology company and cloud provider that withdrew as anchor customer for Nscale's Loughton data center.
xAI
xAI is Elon Musk's artificial intelligence company that installed converted jet-engine turbines to power its Colossus 2 data center.
Crusoe
Crusoe is a Denver-headquartered data center and AI infrastructure company that ordered converted units and abandoned a turbine purchase deal.
ProEnergy
ProEnergy is a company that rebuilds aircraft engines, including a GE engine commonly found on Boeing jets, to burn natural gas.
The details
xAI installed eight converted jet-engine turbines in Mississippi to help power its Colossus 2 data center in Memphis. The converted units come from ProEnergy, which rebuilds a GE engine commonly found on Boeing jets, while Microsoft pulled out as anchor customer for Nscale's data center in Loughton, England.
Timeline
September: Crusoe walked away from a $1.25 billion deal to buy 29 natural gas turbines from Boom Supersonic.
April: Crusoe ordered 13 ProEnergy converted units.
Last month: Microsoft pulled out as anchor customer for the Loughton data center.
2031: Most power capacity is expected to come on by this year.
The Tech Race
The repurposing of GE aircraft engines rebuilt by ProEnergy into stationary gas generators marks a departure from conventional turbine plants built for utilities. Operators are racing to secure power for AI infrastructure as grid timelines stretch toward the early 2030s.
Power constraints on data center construction can influence the cost and pace of AI services that depend on these facilities. Communities near these sites may also be affected by the emissions and efficiency profile of on-site gas generation.
The takeaway
Power availability is increasingly shaping where and how fast AI infrastructure gets built. Watching grid timelines, emissions data, and turbine efficiency will show whether on-site generation becomes a lasting solution or a short-term bridge.
Further reading
For more on how operators are securing power for large facilities, explore Data Centers.
Live Poll
Do you expect your household electricity bills to keep rising in the next few years?
