Microsoft is in exclusive talks with Chevron and investment fund Engine No. 1 over a long-term deal to develop a $7 billion natural gas-fired power plant in West Texas — a move that signals a fundamental shift in how America’s technology industry is approaching the energy problem at the center of artificial intelligence expansion.
The proposed natural gas-fired power plant is projected to cost about $7 billion and initially generate 2,500 megawatts of electricity, making it one of the largest of its kind in the United States, according to people familiar with the matter who asked not to be identified because the talks are ongoing. The facility would be built specifically to power a large data center campus, bypassing the regional electrical grid entirely and supplying electricity directly to Microsoft’s AI infrastructure under a dedicated long-term agreement.
Where the Project Is Being Built and Why Texas
The natural gas-fired facility is located in the Permian Basin, Chevron’s first move into powering AI workloads. This business model allows the company to bypass the regional grid and supply energy directly to a data center operator at a co-located site. The Permian Basin — straddling West Texas and southeastern New Mexico — is the most productive oil and gas region in the United States, giving this project a structural advantage that few other locations in the country can match.
Chevron New Energies is leading development of the West Texas facility, drawing on the company’s natural gas production and operating experience to help meet data center power needs. Division President Jeff Gustavson said the Permian Basin site builds on Chevron’s existing gas production network, giving the project a competitive edge. “In the Permian, we hold a unique data advantage through our interest in one out of every five wells, and we’re building AI-driven tools to unlock that potential,” Gustavson said.
West Texas has become an ideal location because of its energy density. The natural gas produced as a byproduct of oil extraction in the Permian Basin is so abundant that it frequently exceeds pipeline transportation capacity, making the region a natural fit for on-site power generation at scale.
Who Engine No. 1 Is and What Its Role Signals
Engine No. 1 is an activist investment firm that gained widespread attention in 2021 when it successfully challenged ExxonMobil’s board over climate and capital allocation strategy. Its role in this deal is notable: an investor known for pushing energy majors toward discipline and long-term thinking is now positioned as a co-architect of AI infrastructure finance.
Chevron and Engine No. 1 had already announced a partnership to build natural gas-based power plants next to data centers in the United States, with the two planning to use turbines by electric services company GE Vernova. The Microsoft exclusivity agreement now gives that existing partnership a named anchor customer — a development that could accelerate the final investment decision and define the construction timeline.
The Timeline and Scale of the Buildout
With exclusivity talks underway, the partners are moving toward a final investment decision. If all goes according to plan, construction could begin shortly, with the first turbines being delivered in late 2026. This would allow for initial power generation and data center operations to commence by 2027.
The plant’s initial capacity of 2,500 megawatts is already substantial, but the scope may not stop there. The joint venture has mentioned the potential for multi-gigawatt scale, possibly reaching 4 gigawatts or more. If successful, this power foundry model will likely be exported to other regions where energy is abundant but the grid is constrained.
Microsoft has also agreed to rent a data center project in Texas that was originally being developed for Oracle and OpenAI, Bloomberg News reported last week, suggesting the company is assembling Texas-based AI infrastructure across multiple parallel tracks simultaneously.
Why This Deal Reframes the AI Infrastructure Story
The bottleneck constraining artificial intelligence deployment in the United States has changed. For several years, the primary constraint was semiconductors — the GPUs and accelerators needed to train and run large language models. That constraint eased as chip production scaled. The new constraint is electricity, and the Microsoft-Chevron-Engine No. 1 deal is one of the clearest examples yet of how hyperscalers are responding.
Across the industry, developers and hyperscalers are discovering that the biggest obstacle to deploying AI infrastructure is no longer capital, land, or connectivity. It is electricity. In major markets from Northern Virginia to Texas, grid interconnection timelines are stretching out for years as utilities struggle to keep pace with a surge in large-load requests from AI-driven infrastructure.
Morgan Stanley Research forecasts that U.S. data center demand could reach 74 gigawatts by 2028, with a projected shortfall of about 49 gigawatts in available power access. Large technology companies are likely to commit more than $1 trillion of spending in just the 2025–2026 period.
Microsoft is projected to spend up to $146 billion on AI capital expenditures in its 2026 fiscal year. The West Texas plant, at $7 billion, represents a fraction of that annual commitment — but what it represents strategically is something different. It represents a tech giant moving upstream into energy production, treating power generation as a core infrastructure requirement rather than a utility bill.
The Broader Pattern: Energy Companies as AI Partners
This deal does not exist in isolation. Across the industry, technology companies are pursuing similar arrangements with energy producers, moving away from dependence on regional power grids and toward purpose-built generation assets located adjacent to compute campuses.
Electricity consumption in accelerated servers, which is mainly driven by AI adoption, is projected to grow by 30 percent annually in the Base Case scenario, while conventional server electricity consumption grows at a slower pace. Global electricity consumption for data centers is projected to double to reach around 945 terawatt-hours by 2030, according to the International Energy Agency.
Because the existing grid cannot support this surge quickly enough, on-site generation and bridging power solutions are becoming necessary components of the deployment strategy for many organizations undergoing this transition. 2026 marks the beginning of the electrification process; it will continue and ultimately shape the next decade of infrastructure growth, according to industry analysts.
The Microsoft deal with Chevron and Engine No. 1 is a concrete example of that electrification process taking shape at commercial scale. The parties are negotiating a long-term power purchase commitment that will underpin the construction of one of the largest natural gas plants in the country, purpose-built not for the general grid but for a single technology company’s data center campus.
What Happens Next
Exclusivity talks do not guarantee a final agreement. The parties are working toward a final investment decision, and the specifics of the long-term power agreement — including pricing structure, contract duration, and construction phasing — have not been made public. Microsoft, Chevron, and Engine No. 1 have not commented publicly on the negotiations.
If the deal closes as reported, it will mark a milestone in the convergence of American energy and technology sectors — and a data point that will be closely tracked by every other hyperscaler looking at how to secure reliable, scalable electricity for the AI infrastructure decade ahead.
Financial Disclaimer: This article is for informational purposes only and does not constitute financial, investment, or legal advice. All forward-looking statements, projections, and deal terms discussed herein are based on publicly available reporting and unconfirmed sources at the time of publication. Talks described are ongoing and may not result in a completed transaction. Readers should not rely on this article when making financial or investment decisions. Past performance and projected figures are not guarantees of future results. Consult a qualified financial advisor before making any investment decisions.
