Plans to build gas plants respond to increased power demand

Energy companies respond to surging AI demand with behind-the-meter solutions.

Rising power demand, led by the buildout of data centers for Artificial Intelligence (AI), is leading to an increase in natural gas power plant construction. 

Increasing power demand will require significant new power capacity. This is a substantial change after years of stagnant growth in demand and will require large-scale problem-solving. Like most big problems, this is best solved by individual actors driven by the profit motives set out in free markets. As one would expect, energy companies—including those building natural gas—see the importance of AI development and are acting quickly to supply the power that this developing technology needs. 

Some of the new and planned natural gas capacity will be plants that supply the grid at large, but a subset of the new facilities will have behind-the-meter arrangements with data centers. Behind-the-meter arrangements place power plants with sources of large power demand (big industrial customers or data centers) who are then able to use the power. These collocated facilities are then able to use the power “behind the meter” or before it goes onto the broader power grid. These facilities often consume all or most of the power that the plant generates. Because data centers require a large and consistent load, behind-the-meter arrangements often make sense for them, a fact that is unsurprising given that natural gas is the source of more than 40 percent of US electricity. 

One example of new behind-the-meter capacity is a partnership between Chevron, Engine No.1, and GE Vernova to build behind-the-meter natural gas plants to meet data center demand. The initial phase of their plan will supply four gigawatts of power that is targeted to come online by late 2027. 

Engine No.1 Founder and CIO Chris James acknowledged the role that energy plays in the development of AI, stating that “Energy is the key to America’s AI dominance. By using abundant domestic natural gas to generate electricity directly connected to data centers, we can secure AI leadership, drive productivity gains across our economy and restore America’s standing as an industrial superpower. This partnership with Chevron and GE Vernova addresses the biggest energy challenge we face.”

Other companies are working on similar arrangements to build natural gas capacity for data centers, including 360 megawatts planned for the Stargate AI campus in Texas. 

New natural gas capacity, both behind and in front of the meter, is popular right now. Demand for gas plant components is up as more utilities are building new facilities. In addition to its partnership with Chevron, GE Vernova is investing $160 million into its Grenville turbine plant to increase production. This decision will allow the facility to produce 62 turbines in 2026 and 74 in 2027, compared to 37 in 2025 (each turbine is large enough to power 500,000 homes). This is not the only facility in which the company is investing. GE Vernova has $600 million of planned investment across their US facilities. This quick buildout of new production capacity makes sense as electricity demand (and the consequent demand for these components) increases.

As more gas plants are built, the demand for these essential components is rising quickly. Chris White, a GE Vernova spokesman said, “The demand for gas turbines is greater than it’s been in decades… We’re basically booked out for the next three years.”

Gas power plants are a natural solution to rising data center demand because of their dispatchability and comparatively quick construction timelines. As power demand continues to rise from AI and other sources, we are likely to see a continued buildout of new natural gas plants. 

A version of this article first appeared on Catalyst.