The scale of AI infrastructure buildout is staggering with nearly 4,000 data center projects currently under construction globally, representing over $1.2 trillion in value. To power this growth, natural gas turbines are increasingly being deployed as behind-the-meter solutions, with estimates suggesting more than 25% of new facilities above 500 MW will generate their own power this way by 2030, up from just 1% today. High-profile examples include Meta's Louisiana hyperscale campus using H-class turbines, the Oracle/OpenAI Stargate project in Abilene combining gas turbines from GE Vernova and Solar Turbines, and Elon Musk's xAI ordering up to 60 gas turbines for its Memphis supercomputer facility. For those of us in the energy and infrastructure space, this shift signals a fundamental change in how power is sourced, sited, and contracted for large-scale compute.
The explosive growth of hyperscale data centers is driving unprecedented demand for energy, leaving utilities, gas providers, and grid operators racing to keep up. Where is the power coming from currently and where is it likely to come from in the years ahead?
