On Monday, April 20, 2026, Fermi — the nuclear power company building electricity for AI data centers, listed on NASDAQ and the London Stock Exchange under FRMI — abruptly announced that co-founder and CEO Toby Neugebauer and CFO Miles Everson were both stepping down, while unveiling a restructuring plan it calls “Fermi 2.0.” The stock fell 22 percent on the day of the announcement.
The company, co-founded by former US Energy Secretary Rick Perry and Neugebauer, has been one of the most aggressive public bets on the “AI power shortage” thesis: it promises to assemble a 17-gigawatt-class AI campus in Amarillo, Texas, out of natural gas, nuclear, grid power, solar, and batteries. This week, the execution-side pressure burst into the open.
Context matters for why the market reacted so violently. Power delivery has become the binding constraint on new AI capacity; developers that can credibly promise gigawatts on a schedule command premium valuations, and the moment one of them shows governance turbulence, that credibility premium unwinds fast.
Losing a CEO and a CFO in One Day
Per the company’s announcement, Neugebauer resigned as CEO and chairman but keeps his board seat; CFO Everson resigned and was then elected to the board using the director designation rights of the Melissa A. Neugebauer 2020 Trust. Lead Independent Director Marius Haas — former President and Chief Commercial Officer of Dell and founding partner of BayPine — took over as chairman, effective immediately, and stood up an “Office of the CEO”: former COO Jacobo Ortiz Blanes and Anna Bofa, a board advisor since the IPO whose résumé spans Google and Meta, became co-presidents reporting to Haas. A formal CEO search committee of Haas, Lee McIntire, and Cordel Robbin-Coker is working with Heidrick & Struggles; an interim CFO is still being negotiated. The board also elected Jeffrey S. Stein, co-founder of Breakpoint Advisory Partners, as a director.
The arrangement reads as a holding pattern rather than a succession plan: a founder still on the board, an interim structure explicitly billed as providing “stable, experienced leadership” during the transition, and a permanent CFO unsigned. Whatever the press release calls it, the departure of both financial officers at a capital-intensive infrastructure company is the detail professional investors anchored on.
Project Matador: The 17 GW HyperGrid Blueprint
Fermi’s core asset is Project Matador in Amarillo, Texas: a planned 17 GW private “HyperGrid” campus combining one of the nation’s largest combined-cycle natural gas projects, one of America’s largest new clean nuclear complexes, utility grid power, solar, and battery storage — aimed at hyperscaler customers. As part of the reset, the company is moving its corporate headquarters to Dallas and adding an office at the Matador site. Its ties to the Texas Tech University System are billed as a Fermi 2.0 priority; Chancellor Brandon Creighton endorsed the project as having “the potential to deliver generational impact,” while disclosing that the parties are in good-faith discussions to extend certain milestones in the lease agreement.
The Cracks Bloomberg Surfaced
TechCrunch, citing Bloomberg, reports that Project Matador has struggled in recent months, including friction with an unnamed key customer. Read against the announcement’s pledges to “memorialize binding commitments” with the first client-tenants and to court strategic investors — client-tenants and sovereign funds among them — the real job of Fermi 2.0 becomes clear: stopping a trust bleed. A 22 percent single-day drop is the market pricing “governance shock plus customer uncertainty” in real time.
Fermi 2.0: New Faces, New HQ, New Credibility Problem
For the AI infrastructure market, Fermi’s implosion is a timely reality check. The energy side of the AI story has graduated from investor decks to engineering and customer execution, and governance signals, slipping milestones, and customer friction now get amplified straight into the stock. Nuclear-for-AI is not a 2030s narrative anymore; it is a live project pipeline with quarterly disclosure obligations, and that changes what kind of risk can hide and for how long.
For data center and platform teams whose capacity plans hinge on “when does new power actually arrive,” the lesson is concrete: pricing single-power-vendor schedule risk into capacity planning is the same discipline as pricing single-chip or single-cloud risk into architecture — and Fermi’s 22 percent shows the market has started charging for it.
Sources
- CEO and CFO suddenly depart AI nuclear power upstart Fermi — TechCrunch
- Fermi Inc. Announces Fermi 2.0 Strategic Evolution — PR Newswire
- AI nuclear power developer Fermi slumps after abrupt exit of CEO — Bloomberg
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.