The surge in artificial-intelligence computing has collided with a hard physical limit: the electricity needed to run it. In the Texas Panhandle, one developer is answering that demand with a plan of extraordinary scale, pairing a sprawling data campus with its own fleet of nuclear reactors. The plan folds together two of the decade’s defining stories, the explosive rise of artificial intelligence and the return of nuclear power, on a single stretch of Texas ground.
The scale of Project Matador
The development, known as Project Matador, is being built near Amarillo by Fermi America across a footprint of roughly 5,769 acres. The site is designed to deliver up to 11 gigawatts of power, a figure that dwarfs a typical power station and rivals the output needed to light millions of homes, all of it aimed at feeding server halls rather than a city grid.
Alongside that generating capacity, the campus is planned to hold around 18 million square feet of data-center space. The ambition is to keep power generation and computing in the same place, shortening the distance electrons must travel and giving the operator direct control over a supply that AI workloads consume around the clock.
Four AP1000 reactors at the core
The long-term backbone of that power supply is nuclear. According to reporting from DataCenterDynamics, the plan calls for four Westinghouse AP1000 reactors, a modern pressurized-water design that each produce roughly 1,100 megawatts of electricity. That same reactor model is already operating in the United States at plants in Georgia, giving the design a track record that newer, unproven concepts lack.
Reaching that point is a regulated, multi-year process. The developer has moved to finalize the combined operating license application submitted to the Nuclear Regulatory Commission, the single approval that authorizes both construction and eventual operation of a commercial reactor. That review is rigorous and lengthy, and no concrete gets poured on the nuclear side until the commission signs off.
A mixed grid to bridge the gap
Nuclear plants take years to license and build, so the campus is not waiting on reactors alone. The overall design blends several sources, including natural gas, solar, wind, and battery storage, to supply power while the reactors work through licensing and construction. According to the developer’s own description of the campus, that diversified mix is meant to bring capacity online in stages rather than all at once.
Why AI is driving companies toward reactors
The project reflects a broader shift across the technology industry, where the electricity appetite of AI training and inference has outpaced what many regional grids can readily provide. Data centers run continuously and cannot tolerate interruptions, which makes the steady, weather-independent output of nuclear power especially attractive compared with intermittent renewables alone. That reliability, combined with the promise of low-carbon generation, has pushed several operators to consider building or reviving reactors specifically to serve computing loads.
Utilities across the country are now forecasting the steepest growth in electricity demand in a generation, driven in large part by data centers, and several have warned that existing supply cannot keep pace. That backdrop helps explain why a developer would take on the expense and complexity of building reactors rather than simply buying power from the grid.
What the Texas Panhandle offers
The choice of the Amarillo area is not incidental. The Texas Panhandle combines abundant open land, strong wind and solar resources, and proximity to existing energy infrastructure, giving a project of this size room to spread out and multiple ways to generate power. Texas also runs its own electricity grid with a regulatory environment that developers often find faster to work with than those in neighboring regions.
Local and state officials have shown interest in the jobs and investment such a campus could bring to a part of the state better known for ranching and oil than for advanced computing. A build-out measured in thousands of acres and billions of dollars would reshape the regional economy, though it also raises questions about water use, workforce housing, and the strain on local services that any megaproject brings.
The obstacles between plan and power
An announcement of this magnitude is still far from a finished plant. Large nuclear builds have a history of running over budget and behind schedule, and even reactors of a proven design require years of construction after licensing is complete. Financing a project on this scale, securing the skilled workforce, and satisfying environmental and safety reviews all stand between the current blueprints and the first megawatt delivered. For now the campus is a statement of intent, one that captures both the immense energy hunger of modern AI and the lengths developers are willing to go to satisfy it. Even so, the sheer ambition of the proposal signals how far the calculus has shifted, as companies once wary of nuclear timelines now treat reactors as a serious answer to a power problem that shows no sign of easing.
This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.
More from Morning Overview
