Fermi’s Project Matador: Big, Nuclear-Heavy, and Facing a First-Tenant Test
Fermi’s Project Matador is being built as an 11 GW energy and data center campus in the Texas Panhandle near Amarillo, and it’s being billed as the largest AI data center and energy campus on the planet. The data center roof alone will cover about 60 acres, while the full supporting complex stretches across thousands of acres with nuclear units, gas turbines, solar fields, and water treatment. The plan calls for 6 GW of nuclear and 5 GW of gas-fired generation, plus storage and 2.6 million square feet of data-center capacity.
The developers project ultimate compute could reach 20 to 30 exaflops on then-current NVIDIA GPUs, a jump that would let machines perform quintillions of floating-point operations per second. That’s the kind of raw power proponents say will keep the U.S. competitive in AI without shipping load to adversaries. For all of those “flops”, wouldn’t it be a colossal flop if Fermi’s Project Matador can’t sign a anchor tenant? ⁃ Patrick Wood, Editor.
Project Matador is structured as a REIT, so tenant revenue is the mechanism that turns construction into returns for shareholders. That setup makes the identity and timing of the first major tenant critical, and a roughly 21-day delay in formalizing the lease has investors twitchy. Management says the prospective tenant is investment-grade, has signed a letter of intent for a 20-year term plus optional extensions, and advanced $150 million toward initial construction.
CEO Toby Neugebauer has framed the short delay as procedural rather than fatal. “I wouldn’t take anything as a sign of weakness with that delay,” Neugebauer told Wall Street analysts. He added that large corporations have many internal sign-offs that sometimes stretch commercial timelines.
Regulatory progress has been the project’s strong suit, with acceptance of initial sections of a combined license application for AP1000 reactors and active engagement with the NRC on review timelines. Fermi also says it signed a front-end engineering contract for nuclear work and has started production of long-lead nuclear components. Those milestones matter because nuclear capacity is central to the project’s pitch: secure, domestic power for AI at scale.
On the power front, the company expects to begin bringing generation online next year and has regulatory clearance for 6 GW of natural gas-based generation under preliminary approval. Combined-cycle gas turbines and a phased ramp are part of the plan, with one gigawatt targeted to be operational by late 2026. Water supply was also locked down with an agreement for over 2.5 million gallons per day from Amarillo, at a premium rate that shows local buy-in.
There are political overtones built into the project, with the campus dubbed the Donald J. Trump campus and board ties to figures like Rick Perry. The project aligns neatly with a Republican view that America should build AI infrastructure at home and back it with reliable nuclear power. That makes Matador not just a private commercial bet but a piece of broader national strategy to keep critical compute and energy on American soil.
Financials are candidly messy at the moment: the company reported a net loss of $332 million for the first nine months of 2025, which included $173.8 million listed as “charitable contributions” and $111.6 million related to convertible notes and other securities. Investors pushed the stock down after the tenant delay, erasing some of the post-IPO gains despite management beating internal gas-generation estimates. The market is treating the timing of commercialization as the key variable.
The obvious question is why a major data-center customer would hesitate to sign on to a site that promises predictable power and a huge available energy stack. Permitting and power shortfalls have snarled other regions, and Matador markets itself as solving those pain points with integrated generation and a long-term water deal. Supporters argue that once leases are inked, the campus will become a model for linking advanced energy policy to industrial-scale AI capacity.
Matador’s size and ambition mean it will get extra scrutiny, political and commercial, but that same scale is what the Republican case leans on: build big, build fast, and secure America’s AI future with homegrown energy. There’s a gap now between regulatory wins and the revenue-generating leases that convert promises into cash, and closing that gap is the project’s immediate business test.