Fermi, an AI power startup that once commanded a $20 billion market capitalization, has lost both its chief executive and chief financial officer as it attempts to relaunch under a new strategic framework. CEO Toby Neugebauer and CFO Miles Everson have stepped down from their operational roles, though both will retain seats on the company's board. The departures come as Fermi announces what it calls a "2.0" reset — an effort to reposition the company and attract customers for a large-scale data center project planned in Texas.
The leadership shake-up follows a punishing decline in Fermi's valuation. Since going public in a pre-revenue IPO last year, the company's market cap has fallen from $20 billion to roughly $3.4 billion — a drop of more than 80 percent. The slide accelerated after Fermi lost its first hyperscaler customer, a blow that undercut the core commercial premise behind its public listing. A public confrontation between Neugebauer and Howard Lutnick added reputational turbulence to an already fragile situation. Fermi counts former U.S. Energy Secretary Rick Perry among its backers, a connection that initially lent the venture political credibility in the energy-adjacent AI infrastructure space.
The pre-revenue IPO problem
Fermi's trajectory illustrates a recurring tension in the current wave of AI infrastructure investment. Over the past two years, a cohort of startups has gone public on the promise that surging demand for AI compute would translate into enormous energy and data center contracts. Several of these companies listed before generating meaningful revenue, relying instead on letters of intent, memoranda of understanding, or early-stage partnerships with hyperscale cloud providers to justify lofty valuations.
The model works as long as those anchor customers materialize. When they do not — or when a single large contract falls through — the valuation thesis collapses quickly. Fermi's experience echoes patterns seen in earlier infrastructure booms, from the fiber-optic buildout of the late 1990s to the clean-energy SPAC wave of 2020–2021, where capital flowed freely into asset-heavy ventures on forward-looking projections that proved difficult to meet. The loss of a hyperscaler customer is particularly damaging for a company at Fermi's stage because it removes the demand signal that other potential customers and investors use as a reference point.
Retaining Neugebauer and Everson on the board while removing them from day-to-day operations is a common governance maneuver in these situations. It signals continuity to existing shareholders and preserves institutional knowledge while allowing a new management team to negotiate with prospective customers unburdened by prior controversies. Whether the arrangement provides genuine strategic benefit or merely delays a more thorough reckoning depends on who fills the operational vacuum and what terms the "2.0" reset actually entails.
What the reset must solve
The central challenge for Fermi is not merely one of leadership but of commercial credibility. A massive data center project in Texas requires not only capital but binding offtake agreements — contracts in which customers commit to purchasing power or leasing capacity over extended periods. Without such commitments, financing large-scale infrastructure is difficult, and the project risks remaining a planning-stage asset rather than a revenue-generating one.
Texas offers certain structural advantages for data center development: a deregulated electricity market, relatively abundant land, and proximity to natural gas supply. But competition for hyperscaler contracts in the state is intensifying, with established players and well-capitalized newcomers vying for the same pool of potential tenants. Fermi's diminished valuation may complicate its ability to raise additional capital on favorable terms, creating a feedback loop in which the company needs a major customer win to restore confidence but lacks the financial standing to compete aggressively for one.
The broader question is whether the AI infrastructure buildout can sustain the number of ventures currently pursuing it. Demand for AI compute continues to grow, but it is concentrating among a small number of hyperscale buyers with significant bargaining power. Startups that cannot secure anchor tenants face a stark choice between restructuring and dissolution. Fermi's "2.0" framing suggests its leadership believes the underlying asset — the Texas project — retains value if paired with the right team and commercial strategy. The market, having already repriced the company by more than 80 percent, appears less certain.
With reporting from Fortune.
Source · Fortune
