AI Data Centers Turn to Used Jet Engines to Bridge Power Supply Gap
AI firms are buying refurbished jet‑engine turbines to meet soaring electricity needs as new gas turbine deliveries face multi‑year delays.
Key takeaways
- AI data‑center operators are purchasing used jet‑engine‑derived turbines to secure power faster than waiting for new equipment.
- Global gas turbine orders have surpassed 100 GW, outstripping the 60‑70 GW annual manufacturing capacity of major OEMs.
- Used turbines cost roughly $1.5‑$1.7 million per megawatt, offering a discount of up to $2.5 million compared with new units.
AI‑focused data‑center developers are increasingly sourcing second‑hand gas turbines and retired aircraft engines to power their facilities, according to a Decrypt report. The move aims to sidestep a six‑year average delivery lag for new turbines, a bottleneck highlighted by energy consultancy Wood Mackenzie. With global turbine orders topping 100 GW—well beyond the 60‑70 GW annual capacity of manufacturers such as Siemens Energy, GE Vernova and Mitsubishi Heavy Industries—AI firms are seeking faster, cheaper alternatives to meet rapidly rising electricity demand.
Crusoe Leads the Push for Refurbished Power Units
AI infrastructure company Crusoe placed an order for 13 aeroderivative turbines through Missouri‑based ProEnergy, each rated at 50 MW for a total of 650 MW. While the contract does not confirm that the units are salvaged from retired aircraft, the turbines are derived from jet‑engine technology and are intended for future U.S. data‑center sites. The deal exemplifies a broader trend of developers buying refurbished machines that can be delivered within months rather than years.
Cost and Market Dynamics of the Secondary Turbine Market
Used turbines in good condition are selling for about $1.5‑$1.7 million per megawatt, versus $2‑$2.5 million for brand‑new units, according to equipment brokers cited by Reuters. A broker noted a recent sale of two smaller turbines totaling 29 MW for $12.5 million—a price roughly one‑third of the cost for comparable new equipment, which would also have required a five‑year lead time. EthosEnergy CEO Ana Amicarella told Reuters that the secondary market is "very, very hot" as more AI operators compete for limited inventory, driving up asking prices.
Regulatory and Environmental Considerations
Deploying on‑site turbines introduces additional layers of compliance. The U.S. Environmental Protection Agency subjects stationary turbines and engines used by data centers to federal air‑pollution standards, meaning developers must secure permits and manage emissions. OpenAI recently pledged to cover electricity and grid costs for its upcoming campuses after voter‑approved restrictions on data‑center construction, underscoring the regulatory pressure on the sector. Cleanview’s mid‑2026 analysis identified roughly 90 GW of proposed on‑site generation at U.S. data centers, though only about 2 GW was operational at the time.
What to Watch Next
Industry observers will monitor the pace at which used turbine inventories are exhausted and whether manufacturers can accelerate new‑unit production to meet AI‑driven demand. The evolution of alternative on‑site generation, such as gas‑engine generators from firms like Wärtsilä, will also shape the power strategy of data‑center developers. Finally, regulatory developments around emissions and permitting could influence the economics of on‑site turbine deployments.

Discussion 0