But on Wednesday, Fermi outperformed its own goals for its unusually early debut as a public company, highlighting just how much Wall Street hype there is for artificial intelligence ventures.
Fermi's stock closed Wednesday at $32.53 per share, 55% higher than its initial public offering of $21 per share.
That puts Fermi's market capitalization, or its value on the stock market, at more than $19 billion — a remarkable figure for a company that has yet to bring in any revenue and doesn't expect to until at least 2026.
"The team at Fermi America is humbled and pleased by today's IPO performance and already back at work delivering the world's largest next-gen private grid to power next-gen AI, for our stakeholders and the future of our country," spokesperson Cathy Landtroop said in a Wednesday statement.
Dale Klein, a University of Texas at Austin professor and a former federal regulator for nuclear projects, said there's much excitement around AI's potentially huge appetite for electricity and the lucrative opportunities for companies such as Fermi to fulfill that demand.
But he cautioned that Fermi has just started executing its goal to bring its first nuclear reactor online by 2032 — a risk that the company itself acknowledged in its Wednesday filing to federal regulators.
"In the nuclear business, which I'm in, it has a history of overpromising and underdelivering. So when you get on a timeline that compressed, it tends to make you wonder," Klein said.
Powering the future of AI
Fermi burst onto the scene in June, announcing ambitious plans to build what it says will be the world's largest data center campus and the country's largest nuclear power complex near Amarillo, on land owned by Texas Tech University.
If that wasn't headline-grabbing enough, Fermi teased in its initial application to federal regulators that the project — plus the four nuclear reactors planned for the site — could be named after President Donald Trump.
Fermi is moving quickly, Perry said in the company's June announcement, because America is already "behind" other countries in supplying the immense amounts of electricity that could be needed in the race to develop AI.
"The Chinese are building 22 nuclear reactors today to power the future of AI. America has none," said Perry, who served two years as Energy Secretary under Trump's first term.
Investors in Fermi's stock, meanwhile, are betting that tech companies will pay lots of money to host their AI supercomputers at the start-up's yet-to-be-built data center campus. Nuclear is an especially attractive power source because it provides a continuous source of electricity, generated without emissions that cause climate change.
"There's a huge amount of money that is looking for a place to get into nuclear," said Nick Touran, a long-time nuclear engineer and advocate, about Fermi's public debut. "It's not a sure thing that they're going to be able to deliver nuclear, but the money that's rushing to them, it's anticipatory. It's saying, yes, we believe in this idea."
No guarantees
The question is whether Fermi can overcome the challenges that have for decades bogged down efforts to develop nuclear projects in the United States, which have given the industry a reputation for cost overruns and project delays.
There's numerous risks ahead, Fermi said in its filing to the Securities and Exchange Commission Wednesday. For one, the company has no operating history, so "it is difficult to predict what results we might ultimately achieve."
"We have not yet constructed our facilities or entered into any binding contracts with any tenants, and there is no guarantee that we will be able to do so in the future," Fermi said.
Fermi reported a net loss of $6.4 million since its formation in mid-January through the end of June. The company said it doesn't expect to generate any revenue until the first subleases of its data center campus and the delivery of power to those customers.
Since nuclear reactors take years to build, Fermi said its initial plans are to bring 1.1 gigawatts of electricity generation online by the end of 2026, primarily natural gas power plants complemented by solar arrays. That's equivalent to the electricity usage of approximately 275,000 Texas households.
"We expect to incur significant expenses and continuing financial losses at least until Project Matador becomes commercially viable, which may never occur," Fermi wrote in its Wednesday filing, referring to another name for its project near Amarillo.
Filling a vacuum
Still, Fermi CEO Toby Neugebauer told the Houston Chronicle in June that the company is confident it can deliver on its goals because it's hired people who've led successful nuclear projects in China and in the United Arab Emirates.
Additionally, the type of nuclear project Fermi plans to build, Westinghouse AP1000 nuclear reactors, is a tried-and-true technology, unlike some of the other prototypes generating a lot of buzz in the industry.
The challenge is, the last time Westinghouse AP1000 reactors were completed in the U.S. was in 2023 by the Georgia utility Southern Company, an endeavor that was seven years late and $17 billion more expensive than promised. Much of the cost has fallen on the utility's customers.
The Georgia "boondoggle" made it more difficult to finance nuclear megaprojects, even though now is the best time to apply the lessons learned, Touran said. The industry has been trying to figure out how to raise money to try again ever since.
Fermi has provided a new approach: Go public.
"We know the utilities just can't stomach it at this point, so now there's a vacuum. So, how do we get it financed? Here's an answer in Fermi America," Touran said.
© 2025 the Houston Chronicle. Distributed by Tribune Content Agency, LLC.
