AI data centers bypass grid delays by using converted jet engines

Due to protracted waits for utility grid connections and new turbine installations, some artificial intelligence data centers are resorting to using repurposed jet engines to generate their own power. Rystad Energy analyst Audun Martinsen noted that these units allow a data center to become operational in as little as 24 months, contrasting sharply with the International Energy Agency's estimate of a seven-year wait to hook up 100 megawatts in Northern Virginia. Major players are adopting this workaround. Elon Musk's xAI has installed eight converted jet-engine turbines in Mississippi for its Colossus 2 data center, utilizing units from ProEnergy, which rebuilds GE engines from Boeing jets. OpenAI followed this trend in 2025, deploying gas turbines at its Stargate site in Abilene, Texas, while Crusoe, a Denver-based builder, ordered 13 such units in April. These self-generation efforts are becoming critical as delays cause business losses; for instance, Microsoft withdrew as the anchor customer for Nscale's Loughton, England facility after the local utility could not guarantee power until the early-to-mid 2030s. While the Trump administration is encouraging this self-sufficiency, experts caution that these engines are less efficient and more polluting than conventional gas turbines, suggesting they may only be a temporary solution.