US Data Centers Set to Spike Natural Gas Demand

A new report from BloombergNEF reveals that the rapid expansion of artificial intelligence is driving an unprecedented surge in energy requirements. By 2035, data centers in the United States are projected to consume approximately 18 billion cubic feet of natural gas per day. This updated forecast is nearly double what the organization predicted just nine months ago, even after accounting for the likelihood that some announced facilities will never be completed. This massive appetite means data centers will become the second-largest driver of natural gas demand growth over the next decade, trailing only liquefied natural gas exports.

Major technology companies are increasingly taking energy generation into their own hands. Meta, Microsoft, Google, and Amazon have all announced initiatives to build dedicated, onsite natural gas power plants that bypass the traditional electrical grid. BloombergNEF estimates these onsite projects will consume between 2.9 billion and 3.4 billion cubic feet of gas per day by 2035, which equals the total gas consumption of all data centers today. However, the vast majority of the growth will come from grid-connected facilities, which are expected to drive an additional 15 billion cubic feet per day in gas consumption—five times the growth of all other grid-connected sectors combined.

This soaring demand could have severe economic and environmental consequences. Analysts at Noreva warn that the combined pressure of data center expansion and export demands could cause natural gas prices to skyrocket, impacting ordinary utility ratepayers. Furthermore, the environmental toll is staggering. According to the International Energy Agency, burning one cubic foot of natural gas releases about 60 grams of carbon dioxide equivalent. The projected data center demand will add 1 million metric tons of greenhouse gas pollution daily, representing roughly 12% of current U.S. greenhouse gas emissions. For AI practitioners and developers, these figures highlight a growing tension between computational progress and environmental sustainability, suggesting that energy efficiency and alternative power sourcing will soon become critical operational constraints.