AI: US AI Data Center boom driven on natural gas power. RTZ #984

...unlike China leading with nuclear & solar/wind renewables

The Bigger Picture, Sunday, February 1, 2026

In the race to multiple ‘Bragawatt’ AI Data Centers this AI Tech Wave at a cost of hundreds of billions a year, it’s been a given that power sources from gas to nuclear would be needed by the mega-bucket. New data now shows how much of the power for these data centers is scheduled to come from natural gas. That is the Bigger Picture I’d like to unpack this Sunday.

Reminder again that a gigawatt data center today, equipped with power, AI GPUs from Nvidia (most of the time), cooling and other knick knacks runs about $50 billion. That could go down to $30 billion per gigawatt in half a decade or so with AI Infrastructure efficiencies. Maybe.

Most of the ‘Mag 7s’, from Meta, to Google, to Microsoft, to Amazon, to Elon’s xAI/Tesla/SpaceX, and of course OpenAI, are down for hundreds of billions of those tabs EACH, over the next few years. I wrote about the latest budgets this past week alone after the earning results from some of them.

Wired sums up the latest numbers in “Data Centers Are Driving a US Gas Boom”:

“Gas projects in the US pipeline explicitly linked to data centers increased by almost 25 times over the past two years, according to new research from Global Energy Monitor.”

“Data centers have caused the demand for gas-fired power in the US to explode over the past two years, according to new research released Wednesday. More than a third of this new demand, the research found, is explicitly linked to gas projects that will power data centers—the equivalent of energy that would power tens of millions of US homes.”

Of course regulatory relief at the federal level, despite issues at the state and local levels, are providing some tail winds:

“The findings from Global Energy Monitor, a San Francisco–based nonprofit that tracks oil and gas developments, come as the Trump administration is both encouraging data center build-out and doing away with pollution regulations on power plants and oil and gas extraction. They will also almost certainly mean an increase in US greenhouse gas emissions, even if some of the projects tracked by Global Energy Monitor never get built.”

Especially since the US, unlike China, is trying to do it with as little as possible help from renewables like solar and wind. China, on the other hand is doing the opposite, and leading the US in terms of the power race for AI.

“Building all the gas-fired power infrastructure that was in development at the end of last year could increase the US gas fleet by nearly 50 percent, according to Global Energy Monitor’s findings. The US currently has around 565 gigawatts of gas-fired power on the grid. If all the projects in the development pipeline are built, it would add almost 252 gigawatts of gas power to the US fleet. (Estimates vary, but 1 gigawatt can power up to a million homes, depending on the energy use of the region.)”

“Data centers have helped to nearly triple the demand for gas-fired power in the US over the past two years. When Global Energy Monitor last released its tracker, in early 2024, it logged around 85 gigawatts of gas-fired power in the development pipeline in the US. Just over 4 gigawatts of that development were explicitly earmarked for data centers. But in 2025, more than 97 gigawatts of demand tracked were from projects that will be used to power data centers—almost 25 times higher than the 2024 figures.”

“About a year and a half ago, we started to see this increase in proposals for data centers specifically,” says Jenny Martos, a research analyst at Global Energy Monitor who worked on the report.”

The race for power is the key determinant of actually turning on the AI data centers that are being assembled across the country:

“As the data center build-out continues across the country, developers are scrambling to secure power from any and all sources—and utilities are racing to meet the projected demand. This has meant that dirtier power sources are getting a second shot at staying online: coal-fired power plants around the country have recently been given extensions on their retirement dates, boosted by coal-friendly policies from the Trump administration.”

“Natural gas is a much cleaner power option than coal-fired power, but gas plants do release CO₂ emissions. About 35 percent of US energy-related CO₂ emissions in 2022 came from burning natural gas.”

“Gas is cleaner when burnt than coal, but when you’re talking about this much gas, you’re talking about a lot of CO₂ associated with it, too,” says Banks.”

For now, climate concerns take a back seat:

“A larger concern with natural gas is methane leaks during the extraction process. Methane stays for a shorter period of time in the atmosphere than CO₂, but it is 80 times more potent over a 20 year period. Climate scientists say that decreasing methane emissions over the shorter term is crucial to controlling climate change in the long run. It’s estimated that oil and gas production accounts for a third of all global methane leaks; the US is the largest producer of natural gas in the world.”

““If you zero out the methane, and you have a very efficient natural-gas-fired power plant, it is dramatically cleaner than a coal-fired power plant,” says Banks. “But when you layer on the methane emissions, that’s when you see those gaps begin to close between coal and natural gas.”

“A well-regulated oil and gas industry that does its best to stop leaks could go a long way in stopping methane emissions. But this new demand for natural gas from data centers is coming right as the Trump administration is rolling back regulations on the industry. This includes extending the deadline for a major regulation that would require oil and gas operators to take several steps to show that they are tracking and stopping methane leaks.”

“The figures collected by Global Energy Monitor include potential gas units added to a larger grid in order to bolster electricity supply in regions seeing increased demand from data centers. But they also include on-site gas turbines that will solely be used to power data centers. Building power sources on-site, like gas turbines or solar arrays, is becoming increasingly popular for data center developers, who, in some places, are facing yearslong waits to connect to the larger power grid.”

“It’s unlikely that all of the projects tracked by Global Energy Monitor will actually be built. Many data center developers shop around for power with multiple utilities, which helps to inflate overall demand; simply because a data center project has been announced does not mean that it will eventually be built. Efficiency improvements in both data center construction and in AI training will likely decrease some of the projected demand for energy over the next couple of years. A worldwide shortage of gas turbines also will play a big role in determining what actually gets built on time: Two-thirds of the projects tracked worldwide by the Global Energy Monitor do not yet have turbine manufacturers attached to them.”

“We’re in that phase where we’re seeing the explosion in proposals,” Martos says. “What materializes is yet to be determined.”

“Even just a partial build-out of these facilities would be significant. The number of projects under construction—those most likely to be completed—tracked by Global Energy Monitor would add just under 30 gigawatts of gas-fired power to the grid. An additional 159 gigawatts are in the preconstruction phase—not being currently built, but involved in planning and financing processes.”

“AI is not going away—we know that,” says Banks. “It’s really a question of what can we do to diminish the impact of these facilities.”

The whole piece is worth a full read for additional details.

But it’s clear for now that the US is leaning on natural gas as its primary power source for its multi-gigawatt AI data center builds. Unlike China, which is leaning also on renewables like solar and wind, as well as a very well executed plan for nuclear power in their mix.

That is the Bigger Picture on the power for AI data centers in the fourth year of this AI Tech Wave. Stay tuned.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)