AI Data Centers Turn to Retired Jet Engines to Bypass Grid Queues
Developers repurpose aeroderivative turbines as utility connection delays stretch past 2030
Faced with multi-year waiting times for electric utility grid connections, AI infrastructure developers are deploying converted aircraft jet engines to generate on-site electricity. The improvised aeroderivative generators allow hyperscalers to bring energy-intensive compute clusters online years ahead of standard utility interconnection schedules.
Key details
A report by TechCrunch highlights a growing trend among AI cluster builders turning to decommissioned passenger jet engines rebuilt by specialized engineering firms like ProEnergy. These aeroderivative units repurpose General Electric engines commonly salvaged from retired Boeing airliners and convert them to run on natural gas:
- Connecting 100 MW of data center power to the regional grid in major hubs like Northern Virginia now requires up to seven years, according to the International Energy Agency (IEA).
- Major power generation equipment manufacturers, such as Mitsubishi Power, report backlogs stretching order fulfillment for new conventional utility-scale gas turbines out to 2031.
- In contrast, installing repurposed aeroderivative jet turbines enables data center developers to deploy operational on-site power generation within approximately 24 months, according to energy analytics firm Rystad Energy.
- High-profile deployments have accelerated rapidly: Elon Musk's xAI installed eight converted jet-engine turbines in Mississippi to power its Colossus 2 data center in Memphis, while Crusoe ordered 13 converted aeroderivative units from ProEnergy for OpenAI's Stargate data center campus in Abilene, Texas.
Why this matters
The reliance on salvaged aircraft engines illustrates the acute severity of the electric power supply bottleneck facing the artificial intelligence industry. As frontier training runs require hundreds of megawatts of continuous baseload capacity, the speed of hardware delivery has outstripped the expansion capacity of traditional utility grids. Bypassing utility infrastructure through self-generation allows developers to monetize AI accelerators immediately, but aeroderivative engines operate with lower fuel efficiency and emit higher levels of criteria pollutants than modern combined-cycle power plants.
Context
Grid connection logjams have disrupted major infrastructure deals across the globe. Microsoft recently withdrew from an anchor customer agreement for Nscale's planned AI data center facility in Loughton, England, after local utility operators revealed they could not provide necessary power transmission capacity until the 2030s. In the United States, federal initiatives and executive policies have actively urged data center operators to bring their own power generation rather than placing additional strain on community electrical networks.
Risks and open questions
Deploying converted jet engines presents significant environmental and economic compromises. Aeroderivative gas turbines burn more natural gas per megawatt-hour than permanent combined-cycle generation facilities, leading to elevated greenhouse gas emissions and higher operating costs. Furthermore, local air quality regulators and nearby residents have raised objections over nitrogen oxide emissions, noise pollution, and permitting exemptions associated with behind-the-meter generation arrays in industrial corridors.
What happens next
Data center operators view converted aircraft engines as transitional bridge power while waiting for permanent utility transmission upgrades and custom power equipment manufacturing. Regulators at state and federal levels are expected to review whether long-term operation of aeroderivative turbines complies with clean air standards or whether stricter emissions restrictions will be enforced on off-grid AI compute facilities.
Source: TechCrunch Published on AI Usage Global, author: AUG Bot
