AI Boom Drives Shift Toward On-Site Natural Gas Power
Data center developers are deploying on-site natural gas turbines and engines to bypass utility grids and meet massive AI electricity demands.
Data center developers are increasingly implementing on-site natural gas power generation to satisfy the electricity requirements of the artificial intelligence boom. By using behind-the-meter solutions, these companies bypass traditional utilities to ensure a stable power supply for high-demand compute clusters.
Titus Low Carbon Ventures is deploying reciprocating engines for baseload power in Texas data center parks, arguing that these engines are superior for managing transient loads. Other major initiatives include a $1.25 billion deal where Boom Supersonic will provide 29 jet-engine gas turbines to Robinson Crusoe for data centers supporting OpenAI. In El Paso, Texas, Meta has deployed more than 800 mobile mini-turbines to maintain operations.
Texas has emerged as the center of this trend, with nearly 58 gigawatts of natural gas power currently in development. Additional large-scale projects include BorderPlex's proposed $165 billion Project Jupiter campus in New Mexico, which intends to use simple-cycle gas turbines, and facilities in West Virginia supplied by Caterpillar engines.
Environmental researchers and advocacy groups warn that these developments lock in years of fossil fuel reliance. They estimate this shift could add 44 million metric tons of carbon dioxide to the atmosphere by 2030 and hinder the adoption of clean energy alternatives.