Rick Perry’s Fermi Is Undermining the AI Energy Thesis
(Bloomberg Opinion) — The Fermi Paradox asks why, given the vastness of the universe, there is no concrete evidence of aliens. Fermi Inc. The paradox asks why, given hyperscalers’ seemingly insatiable demand for electricity, former Energy Secretary Rick Perry’s AI energy firm hasn’t received hard power contracts from said hyperscalers.
The dissonance has become unbearable. Fermi had a good weekend, with shares down 69% since its initial public offering last September. Chief Executive Officer, co-founder and major shareholder Toby Neugebauer left suddenly. Then Chief Financial Officer Miles Everson followed him to the door. On Monday morning, the company was announcing a new “CEO office” consisting of a board member and chief operating officer. Details of the new interim CFO will be announced later this week.
The stock is down another 20%, which means a dollar and change at this point. Fermi’s turmoil is an alarm bell for the overhyped hyperscaler thesis, which has boosted industry-wide power estimates and valuations.
The revolution of the weekend’s top executives was completely constructed as “Fermi 2.0”. But just six months after its sensational stock market debut, a company that updates itself faster than Microsoft software patches is not a good sign. Fermi’s announcement of a search for an “interim” CFO also adds an air of unhelpful improvisation to the proceedings. The appointment of Jeffrey Stein, who co-founded the restructuring of boutique Breakpoint Partners and has “more than 30 years of experience in the distressed and special situations market,” to the board is also something no investor likes to hear on a Monday morning.
As I wrote last October, Fermi had “all the trappings of an IPO.” Private equity mogul and Republican donor Neugebauer’s name, as well as Perry’s co-founder, also attracted attention. The company had secured a lease for a strategically located field in the Texas panhandle, as well as a gas supply contract and orders for hard-to-get gas turbines. This and Fermi’s ambitious ambition to build nuclear reactors made the site near Amarillo seemingly perfect for any hyperscaler looking to get multi-gigawatt data centers up and running quickly. A small public offering, along with ballooning forecasts for data center power demand, pushed Fermi’s valuation to a peak of more than $19 billion.
This currently hovers around $3 billion. The scene was set on a defensive earnings call late last month. Investors had become uneasy about the ongoing lack of an anchor tenant, especially after a Business Insider report in December alleging that Amazon.com Inc. had backed out of a potential deal, citing Neugebauer himself as a source. (Fermi later denied that the company or Neugebauer had identified the prospective tenant; Business Insider stood by its report).
Neugebauer and Everson spent most of the meeting discussing possibilities for a tenant agreement and listing financing sources, mostly non-recourse equipment financing loans. Fermi’s total capital expenditure in 2026 and 2027 is estimated to be $6.8 billion; That’s nine times forecast EBITDA and more than double current market capitalization. At one point, Neugebauer said Fermi had the company on a rolling phone call every day at 4 p.m. Eastern to review its cash position; This may have been intended to reassure, but rather it raised the question of why managers feel the need to do this every day. Of course, the former CEO and CFO did not reverse this sentiment.
Here’s a pretty standard cautionary tale about overhyped IPOs. Fair play to whoever came up with “Fermi 2.0,” but its already tenuous persuasive power is further weakened by the continued and awkward presence of Neugebauer and Everson, both now on the board.
But this warning extends far beyond Fermi’s non-insider shareholders. The AI boom and the dizzying amounts of financing that have come with it have fueled a make-make-future mentality in the energy sector. Besides Fermi’s initial stock prices, advanced nuclear reactor developers, most of whom do not have a licensed design, sport meme-like valuations; Oklo Inc., for example, is down nearly 60% from its peak last fall and is still trading at more than 1,200 times forecast revenue.
But as Jigar Shah, former head of the Department of Energy’s Office of Credit Programs who now runs the consulting firm Multiplier, wrote in a timely thread on X this weekend: “This is not a financing issue.” Predictions of a 100-plus gigawatt increase in power demand from Silicon Valley executives more accustomed to the infinitely scalable dynamics of software are running up against the stubborn physical realities of the grid (along with a growing political backlash). The supply chain and regulatory process for power connections is redundant and subject to broader stakeholder groups than just hyperscalers. Shah estimates that about 34 gigawatts of capacity could be added by 2030.
Part of the problem here is that hyperscalers demand a level of reliability that the U.S. electric grid will struggle to provide, and that requires redundancy in the form of on-site or backup generation. Sector and Sovereign Research LLC predicts data centers will add 75 gigawatts of peak demand by 2030; this is the lion’s share of the total increase, which they estimate at 105 gigawatts. However, the need to maintain reserve capacity buffers both on and off the grid means that the net capacity required will be 182 gigawatts. “The United States has never added 182 GW of dispatchable capacity in a five-year period,” SSR writes.
Fermi builds on this and actually provides a shortcut to connecting to hyperscalers. But an AI industry seemingly desperate for power and rich enough to pay is strangely hesitant. The more enthusiastic forecasters of data center demand should surely pause in the face of such evidence, or rather the lack of it.
More from Bloomberg Opinion:
This column reflects the author’s personal views and do not necessarily reflect the views of the editorial board or Bloomberg LP and its owners.
Liam Denning is a Bloomberg Opinion energy columnist. A former banker, he edited the Wall Street Journal’s Heard on the Street column and wrote the Financial Times’ Lex column.
More stories like this available Bloomberg.com/opinion
