Soaring AI Demand Has Data Centers Deploying Jet Engines To Keep Up
This is the reality of the AI gold rush, where the computational power of a small city is now being delivered not by a new utility plant, but by the same gas turbines that once pushed a Boeing 747 across the Atlantic.
So the gist is that AI data centers are exponentially hungrier than their traditional counterparts. A typical facility augmented with high-performance GPUs can require more than ten times the energy of a conventional server farm. The largest complexes currently being planned could consume up to 2 gigawatts, equivalent to a major nuclear or gas power plant! By 2035, the power demand from these AI-focused facilities in the United States alone is estimated to grow by more than thirty-fold.
Yet, utility companies are dragging their feet (or have been caught with their pants down) on the matter. Data center operators and developers can't wait around for the grid to catch up, so the industry’s pragmatic answer comes in the form of aeroderivative gas turbines—power generators built by retrofitting proven aircraft engines. Machines, like the GE Vernova LM6000 and the ProEnergy PE6000, are lightweight, compact, and—critically—can be deployed far faster than traditional utility-grade plants. They are being installed as power bridges, designed to run the data center for the first five to seven years until the permanent grid connection is ready. ProEnergy sells units (based off of the CF6-80C2 (used on Boeing 747s, 767s, and Airbus A300s, and A310s) capable of producing up to 48 megawatts at a time, in fact.
Nonetheless, in cities like Memphis, Tennessee, Elon Musk's xAI data center facilities have already drawn controversy and local protests for operating gas turbines without appropriate permits, sparking environmental justice concerns in areas already burdened with poor air quality.
