Fermi Fires Co-Founder After $19 Billion Nuclear AI Startup Fails to Sign a Single Client

--- headline: "Fermi Fires Co-Founder After $19 Billion Nuclear AI Startup Fails to Sign a Single Client" slug: fermi-nuclear-ai-fires-cofounder category: business story_number: "03" date: 2026-05-02 author: The Vault AI ---

When Fermi America went public in October 2025, investors tripped over each other to buy shares in the company that promised to marry nuclear power with artificial intelligence at a scale never before attempted. Eight months later, the startup has no clients, an 84 percent stock crash, a fired co-founder, and a growing collection of class-action lawsuits. It is one of the most spectacular flameouts in the short but frenzied history of AI infrastructure investing.

The pitch that lit up Wall Street

Fermi America was founded by Dallas billionaire Toby Neugebauer and former Texas governor and U.S. Energy Secretary Rick Perry. Their pitch was simple and seductive: the AI boom needs power, nuclear is the answer, and Texas is the place to build it. The company planned to construct what it called Project Matador -- a 5,800-acre campus near Amarillo, Texas, branded the Donald J. Trump Advanced Energy and Intelligence Campus. The vision was staggering: 17 gigawatts of planned capacity, roughly three times the peak electricity demand of New York City, generated by a combination of four Westinghouse AP1000 nuclear reactors, natural gas, and solar installations. The campus would house more than 18 million square feet of data centers packed with supercomputers.

At its peak, Fermi commanded a market capitalization approaching $20 billion. Shares traded near $37. Wall Street loved the narrative: a politically connected team, a massive land play in the Texas panhandle, and a seemingly insatiable demand for AI compute.

The unraveling

The problems started becoming impossible to ignore in early 2026. In March, the company reported a full-year 2025 loss of $486.4 million. More damning was the termination of a $150 million Advance in Aid of Construction Agreement by what would have been its first major tenant -- the single financial commitment underpinning the entire first phase of Project Matador.

Then came the leadership exodus. On April 20, Fermi announced the departure of both its CEO, Toby Neugebauer, and its CFO, Miles Everson, in what the company euphemistically described as a "strategic evolution" and a "2.0 reset." The market was not fooled. Shares plunged 18 percent on the news alone, adding to a decline that had already erased the vast majority of the company's value.

By late April, the situation deteriorated further. Bloomberg reported on May 1 that Fermi's board had formally fired co-founder Neugebauer, citing violations of company policies without providing detailed explanations. The stock, which once traded near $37, had cratered to roughly $5.28 -- an 84 percent decline from its peak. Class-action lawsuits quickly followed, with plaintiffs alleging that Fermi misrepresented demand for Project Matador and concealed the risk of relying on a single tenant's funding commitment.

Amid the carnage, investigative reporting revealed that company insiders had been dumping millions of shares even as they publicly talked up the venture's prospects.

Zero clients, zero revenue, big questions

The most devastating detail in the entire saga is the simplest one: Fermi never signed a single paying client. Despite months of pitching hyperscalers and enterprise AI companies, the startup could not convert its grand vision into a single binding contract. The first gigawatt of capacity was supposed to be operational before the end of 2026, with nuclear construction beginning in 2027. Those timelines now look like artifacts from an alternate reality.

Rick Perry's political connections -- once touted as a decisive advantage in navigating nuclear permitting and federal energy policy -- proved insufficient to close the gap between narrative and execution. Land, reactors, and a former cabinet secretary's Rolodex are not, it turns out, a substitute for customers.

What this means for AI infrastructure hype

Fermi's collapse is not an indictment of nuclear power for AI, nor does it invalidate the very real demand for massive new energy sources to fuel data centers. What it does expose is the danger of valuing narrative momentum over commercial traction. Investors poured nearly $20 billion in market value into a company that had not yet demonstrated it could do the most basic thing a business must do: persuade someone to pay for its product.

The AI infrastructure buildout is real. Microsoft, Google, Amazon, and Meta are each spending tens of billions on data centers. The need for clean, reliable baseload power is genuine. But the gap between identifying a macro trend and building a viable business within it remains enormous. Fermi bet that the market would reward ambition and connections. The market, eventually, demanded receipts.

What to watch next

Several questions will determine whether Fermi survives or becomes a cautionary footnote. First, can the reconstituted board find a credible CEO and secure an actual client before cash runs out? Second, will the class-action litigation expose deeper governance failures? Third, and most importantly for the broader market, will Fermi's implosion cool investor appetite for pre-revenue nuclear-AI ventures, or will the next pitch deck simply learn to name-drop different politicians?

For now, the 5,800 acres near Amarillo sit largely empty -- a monument to what happens when hype outpaces execution by $19 billion.