AI Data Centers Turn to Decommissioned Jet Engines as Grid Queues Stretch Years

Facing multi-year waits for grid connections, several major AI companies are buying converted jet engines pulled from retired aircraft to generate their own electricity — a solution that is faster to deploy but less efficient and more polluting than purpose-built turbines.

The primary driver is speed. Connecting 100 megawatts to the grid in Northern Virginia takes about seven years, according to the International Energy Agency [1]. Most new power capacity is not expected online until 2031, a Mitsubishi Power executive told Bloomberg [1]. By contrast, converted jet-engine units can have a data center running in roughly 24 months, Rystad Energy analyst Audun Martinsen told the New York Times [1][3].

"It's all about getting power as fast as possible," Martinsen said [1][3].

The delays are already forcing business decisions. Last month, Microsoft withdrew 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 [1][2][5].

Who is deploying them

Elon Musk's xAI has installed eight converted jet-engine turbines in Mississippi to help power its Colossus 2 data center in Memphis, according to state permits and satellite imagery reviewed by the New York Times [1][3]. The units come from ProEnergy, which rebuilds a General Electric engine commonly found on Boeing jets [1][3][4].

OpenAI was reported to use gas turbines at its first Stargate site in Abilene, Texas in 2025 [1][5], though a later account says Crusoe's 1.2 GW Abilene campus built for Oracle and OpenAI is powered by the grid with gas turbines used only for backup [8]. Crusoe, the Denver-based startup constructing that campus, ordered 13 ProEnergy units in April 2026 [1][5]. It is unclear from the sources whether those units are destined for the OpenAI campus or for Crusoe's separate 900 MW Microsoft campus in Abilene, which the company says will be powered by on-site gas turbines [8].

In September, Crusoe walked away from a $1.25 billion deal to buy 29 natural gas turbines from Boom Supersonic, a startup developing a supersonic passenger jet that had begun selling a version of its engine as a stationary power plant [1][5][8]. A Crusoe spokesperson said the company still plans to use turbines, "just not Boom's" [8].

ProEnergy has emerged as the leading supplier. Each converted turbine produces about 50 megawatts — enough for roughly 15,000 homes [3]. FTAI Aviation, a New York-based jet-engine leasing and repair company, launched a conversion division last year and signed a $1.5 billion agreement in July to supply turbines to a data-center developer [3]. Boom Supersonic says it will deliver about 250 megawatts of its Superpower turbines to other sites in 2027, targeting 1 gigawatt in 2028 [8].

Efficiency and emissions trade-offs

Converted jet engines carry significant compromises. They are less fuel-efficient than turbines designed for stationary power plants and produce higher emissions per megawatt, requiring more units to generate equivalent power [1][2][3][4][5].

State permits for xAI's Mississippi units allow carbon monoxide emissions at twice the rate of other turbines at the site, and a 25 percent higher limit for volatile organic compounds — both posing health risks [3]. Patrick Anderson, an attorney with the Southern Environmental Law Center, said electric generation for AI "should be getting cleaner over time, not dirtier" [3].

Google told the Times it has opted to wait for grid-connected power rather than use converted engines [3].

Policy pressure

The Trump administration has encouraged data centers to generate their own electricity. In an October 21 Truth Social post, Trump said new large-scale data centers "are mandated by me to make their own electricity" and sell excess back to the grid [6]. In a court filing, the administration sided with xAI against an environmental challenge, arguing the lawsuit threatened national security by undermining "global AI dominance" [3].

Industry experts caution that onsite generation faces its own supply-chain constraints. Turbine models are effectively sold out through 2031, and permitting, fuel supply, and maintenance vary widely by location [6]. Patrick Hughes of the National Electrical Manufacturers Association said "there's no single answer, as the right onsite generation strategy will inevitably depend on each facility's load, site, and regulatory landscape" [6]. The DataCenterKnowledge article notes that for many facilities, a hybrid approach combining onsite generation and storage with a grid connection may be the most practical outcome [6].

A temporary bridge?

Analysts describe jet-engine conversions as a short-term play. Efficiency and environmental concerns may grow in importance within a year or two as regulatory scrutiny intensifies [1][2][5]. For now, the economics favor speed: with chips "piling up" and unable to be turned on, as Musk put it in February [3], the fastest path to revenue wins — even if it means burning more gas in engines never designed to run on the ground.