Fermi Inc. lost its chief executive on Friday, its chief financial officer on Sunday, and roughly a fifth of its remaining market value by Monday morning. The company called the leadership overhaul “Fermi 2.0.” Wall Street called it a sell signal.

Shares in the Texas-based power infrastructure company closed at $5.40, down about 18 percent, according to CFO Dive. That brings Fermi’s cumulative decline to 84 percent since its October 1 debut at $32.53 — wiping more than $16 billion from a peak valuation that briefly topped $19 billion. The company now hovers around $3 billion.

A Weekend Coup, Packaged as Strategy

Co-founder and CEO Toby Neugebauer departed on April 17, according to an SEC filing. Two days later, CFO Miles Everson resigned — “without Good Reason,” as the filing noted, meaning he forfeited certain contractual protections. The board installed an “Interim Office of the CEO” comprising chief operating officer Jacobo Ortiz Blanes and board observer Anna Bofa. A replacement CFO is being negotiated, with an announcement expected this week.

Both men remain on the board. Neugebauer continues as a director. Everson’s seat comes via a director designation right exercised by the Melissa A. Neugebauer 2020 Trust, a vehicle linked to Neugebauer’s family. The board had expanded from five to seven seats on April 17, and two days later Jeffrey Stein filled one of the new vacancies, co-founder of Breakpoint Advisory Partners — a boutique corporate advisory and investment banking firm, per his SEC biography.

A restructuring expert joining the board of a company barely six months past its IPO is, to put it gently, an unusual signal.

The Nuclear-AI Pitch, Untested

Fermi was co-founded by Neugebauer and Rick Perry, former Texas governor and US Energy Secretary from 2017 to 2019. Its flagship Project Matador campus near Amarillo, Texas, is supposed to deliver private power grids at the scale AI data centers require — including what Politico called “the most ambitious build-out of legacy nuclear reactors in half a century.”

The pitch was straightforward: hyperscalers like Amazon, Microsoft, and Google need vast quantities of electricity, the US grid cannot deliver it fast enough, and Fermi would offer a plug-in-ready shortcut. That thesis, combined with a tiny free float, propelled the stock to a $19 billion valuation.

Six months later, Fermi has not signed a single anchor tenant.

A December Business Insider report claimed Amazon had pulled out of a potential deal, citing Neugebauer as its source. Fermi denied that either the company or Neugebauer had identified the prospective tenant. Business Insider stood by its reporting, according to Bloomberg Opinion.

The Friction Behind the Ouster

The leadership purge may not be purely strategic. Stifel analyst Stephen Gengaro told clients the company had signaled “friction between customers and the outgoing CEO,” suggesting Neugebauer’s departure could smooth negotiations with hyperscalers, MarketWatch reported.

That is the generous read. The less generous one: a company projecting $6.8 billion in combined capital expenditure over 2026 and 2027 — nine times forecasted EBITDA and more than double its current market capitalization — just lost the two executives responsible for raising the money to fund it.

On a defensive earnings call late last month, Neugebauer told investors that Fermi held a standing call every day at 4 pm Eastern to review its cash position. He presumably meant this as reassurance.

A Warning Beyond Fermi

Fermi’s unraveling carries implications for the entire AI-energy thesis. The conviction that hyperscaler demand would justify building massive power infrastructure on spec has inflated valuations across the sector. Oklo Inc., another nuclear-AI play, has fallen roughly 60 percent from its peak but still trades at more than 1,200 times forecasted revenue, according to Bloomberg Opinion.

The physical realities are catching up. Jigar Shah, former head of the Department of Energy’s Loan Programs Office, wrote on X this weekend: “This isn’t a funding problem.” Supply chains and regulatory processes for grid connections are backed up. Forecasts of 100-plus gigawatts of new data center demand are colliding with the limitations of the US power system.

Sector and Sovereign Research estimates that meeting projected demand through 2030 would require 182 gigawatts of new dispatchable capacity. “The United States has never added 182 GW of dispatchable capacity over a five-year period,” the firm wrote.

Fermi’s backers wagered that hyperscalers would pay whatever it took to jump the line. Six months in, those customers are still not buying. The executives who promised they would are now on the board, no longer running the company. And the market has noticed.

Sources