AI Data Centers Swap Grid Lines for Retired Jet Engines

Waiting seven years for a standard electrical grid connection is no longer an option for companies racing to build out artificial intelligence infrastructure. Some data center operators are skipping the queue entirely by turning to retired aircraft jet engines rebuilt to burn natural gas. This workaround allows facilities to become operational in just two years, offering a stark contrast to the glacial pace of traditional utility hookups.

The urgency is reshaping energy strategies across the industry as power supply desperately lags behind demand. Elon Musk’s xAI deployed eight converted jet-engine turbines in Mississippi to support its Memphis data center, while OpenAI incorporated similar gas turbines at a Texas site managed by startup Crusoe. These modified units, sourced from companies like ProEnergy, provide immediate electricity but introduce significant environmental and operational compromises.

According to reports, the retrofitted engines are less efficient and produce more pollution than turbines built specifically for conventional power plants, requiring higher unit counts to match standard output. Our read is that while these makeshift power plants solve an immediate bottleneck, they risk trading long-term sustainability for short-term velocity.

When major customers like Microsoft pull out of multi-year deals due to utility power gaps well into the next decade, operators will reach for any available stopgap.

That reality makes aircraft engines an attractive temporary fix, but efficiency demands and environmental scrutiny may force a reckoning for these temporary plants within a year or two.