Chevron and activist investor firm Engine No. 1 are teaming up on plans for what they expect will be the first multi-gigawatt gas-fired power plant co-located with a data center.
They said Jan. 28 that their partnership to build a new company to develop scalable and reliable power solutions for U.S. data centers is based on President Donald Trump’s supportive early moves for U.S. energy development, including in support of artificial intelligence.
They said they have secured manufacturing slots for seven of GE Vernova’s 7HA gas turbines and expect to install them at data centers in the Southeast, Midwest and West that they are calling “power foundries.”
They would total as much as 4 GW of capacity and would be targeted to be in service by the end of 2027. There is potential to expand beyond this capacity and potential for future addition of carbon capture and storage or other carbon-reduction strategies.
The generating units initially would not send power through the grid. But the model is designed to allow future interconnects.
Chevron, Engine No. 1 and GE Vernova offered comments directly in line with their roles as an oil supermajor, an industrial investment firm and a power equipment manufacturer.
Chevron CEO Mike Wirth said: “We are proud to play our part in bringing to fruition President Trump’s vision for a new American golden age, powered by our enormous energy resources and unrivaled workforce.”
Engine No. 1 CIO Chris James said: “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.”
GE Vernova CEO Scott Strazik said: “GE Vernova is uniquely positioned to provide the energy systems and support required to make this large-scale endeavor possible, as the leading U.S. energy manufacturer.”
Almost all observers expect U.S. electricity demand to increase in coming years, in part because of the rise of energy-intensive artificial intelligence computing. There has been some disagreement on how sharply it will increase, however.
