Experts are still divided over how much power demand for artificial intelligence (AI) will actually increase daily US natural gas demand in the next decade, despite extensive discussion over the past year.
The proliferation of AI and the data centers that power it may only increase daily demand by 3.3 Bcf (93mn m³), according to Sharon Paterson, the director of business development at MiQ. She made her comments Tuesday at the Midcontinent LDC Gas Forum in Chicago, Illinois.
That estimate is beneath even the lowest range of other estimates made at the same forum. East Daley Analytics senior director Jack Weixel forecasted a 5-7 Bcf/d increase, while ConocoPhillips senior market analyst James Pearson estimated a 5-8 Bcf/d growth.
While all of these analysts acknowledged that AI would meaningfully drive gas demand in the coming years, they also all discussed the difficulty of making predictions with any degree of specificity.
AI demand growth has been hotly debated over the past two years. EQT, the US's largest gas producer by volume, estimated in 2024 that AI would boost daily demand by 10-18 Bcf, while Kinder Morgan forecast a 7-10 Bcf/d growth.
Both those estimates are mostly above what has been forecast at the 2025 LDC forum. More analysts have recently begun to acknowledge how difficult it will be to bring on enough electrical capacity to power the full extent of possible AI demand.
Just one combined-cycle gas turbine (CCGT) power plant came on line last year, at the Plaquemines LNG terminal. Although nearly 19GW of new capacity is planned by 2028, only 4GW is under construction, while the rest is waiting for regulatory approval. A large amount of new electrical capacity will have to be successfully built in order to power the level of demand forecast by some analysts.
Equipment shortages — gas turbines and transformers — and electricity grid constraints also inhibit capacity growth. Some analysts have begun to pare down their demand estimates in order to account for these regulatory and equipment roadblocks, cutting into expectations for rapid demand growth in the next decade.
