AI Data Centers Turn to Used Jet Engines and Turbines as Power Crunch Worsens

The race to power artificial intelligence data centers has grown so intense that developers are bypassing years-long queues for grid connections and new turbines by converting retired jet engines into on-site electricity generators. The engines, pulled from decommissioned aircraft, offer a faster if less efficient and less environmentally friendly path to power, according to reporting cited by Yahoo Finance. The workaround reflects the severity of a bottleneck that has already cost some operators business: Microsoft withdrew as anchor customer from Nscale's Loughton, England data center after the local utility said it could not deliver sufficient power.

The scramble for unconventional power sources is feeding broader debate about whether the massive data center buildout can be justified. Commentary published November 9 argues that unless the tech industry finds a way to make AI agents substantially more useful and popular, the trillion-dollar infrastructure investment amounts to a bad bet. That skepticism is showing up in markets: Bloom Energy, whose solid-oxide fuel cells offer an alternative power solution for data centers, fell roughly 30% following the recent election, though at least one analyst now frames that selloff as an asymmetric buying opportunity given U.S. data center pipeline demand of around 346 gigawatts against Bloom's current deliveries of only 1-2 gigawatts.

The power crunch is intersecting with a broader pullback in AI-related equities. Semiconductor stocks tied to AI have retreated sharply enough that analysts are identifying potential recovery plays, while the contrast between Palantir's profitability and Snowflake's ongoing losses is drawing fresh scrutiny heading into 2027.