Artificial intelligence (AI) is increasingly influencing electricity consumption in the U.S., mainly due to the rapid expansion of data centers. Sandy Segrist at Hart Energy recently wrote an article titled “Behind the Hype: The ‘Jaw-dropping’ Expectations for AI, Natural Gas“. This is a summary of that article. Subscribe to Hart Energy for the complete article and more oil and gas news.
Natural gas remains a favored energy source due to its reliability and lower emissions compared to coal or oil. It is particularly crucial in data centers, which are expected to consume as much as 9.1% of U.S. electricity by 2030, up from 4% in 2024. However, integrating AI data centers into the energy market is challenging. These facilities often require stable, long-term energy contracts, while the natural gas market is subject to significant price volatility. Further complications arise from the short lease durations common in the tech industry, which may conflict with the long-term investments needed for power generation infrastructure.
Renewable energy is also expanding, driven by technological advancements and a push towards sustainability—however, renewables like wind and solar face location, intermittency, and reliability issues. As a result, tech companies and data centers often rely on natural gas as a transitional energy source. The future could see a more diverse energy mix, potentially including nuclear and hydrogen, to “bridge the gap” left by renewables.
Virginia currently leads the U.S. in data center energy demand, accounting for 22% of the total. However, AI data centers are increasingly being developed in regions like Texas and the Dakotas, where renewable and natural gas resources are available. Some analysts caution that the expected surge in natural gas demand may not fully materialize due to efficiency gains and increasing renewable use. The debate continues, but for now, natural gas provides a reliable solution for the growing energy needs of AI infrastructure.
Change in annual electricity sales to commercial customers, gigawatt hours (GWh). (Source: Energy Information Administration)
Original Article HERE
