AI data-center giants risk costly gas price spikes as new forecast warns
A fresh Noreva report cautions that natural-gas prices could triple in parts of the United States, threatening the economics of hyperscalers' new gas-powered AI data centers.
Major cloud providers have begun constructing gigawatt-scale natural-gas power plants in Louisiana and Texas to feed AI-driven data centers, betting on historically low gas prices. Noreva, an energy-research firm, warns that this assumption may prove costly, forecasting that prices in certain U.S. hubs could climb above $10 per million BTU, compared with today’s $2-$4.5 range. The projected surge stems from constrained supply growth, expanding LNG export pipelines, and a new demand pull from AI workloads.
If gas costs double or triple, operating expenses for “bring-your-own-power” facilities could rise sharply, potentially inflating token prices or forcing a shift to grid electricity, which would lift overall power rates. While futures markets currently show stability, the analyst doubts that the market will remain calm. The situation highlights how hyperscalers, traditionally cautious about capital projects, are now entangled in the volatile fossil-fuel sector.
Why it matters
Higher gas costs could make AI data-center operations far more expensive, affecting tech company margins and consumer energy bills.
In this story
