Fermi
Vertically integrated developer-REIT: builds on-site power generation (natural gas near-term; nuclear/solar/behind-the-meter later) plus data-center shells on a single Texas campus, intended to be monetized via long-term power + colocation leases to hyperscalers/neoclouds. No operating leases signed yet; revenue is contingent on landing a binding anchor tenant.
Earnings, margins, COGS & capex
Development-stage company with no revenue and rapid cash burn. Q1 2026 net loss $189M (~70% non-cash: ~$134M share-based comp plus a $25M debt-extinguishment loss); ~$441M deployed into the campus in the quarter, lifting gross PP&E to ~$1.43B. Liquidity is thin (~$243M cash incl. $35.8M restricted) against ~$421M debt, and the model is structurally dependent on securing a binding hyperscaler tenant to unlock construction financing -- a dependency thrown into doubt when a ~$150M anchor arrangement was terminated in Dec 2025 and no replacement had been signed by the March target.
Revenue trend
Margins
no revenue
large losses expected to persist through construction
worsening with build-out pace
COGS structure
No cost of goods sold yet (no revenue). Future COGS would be fuel (natural gas), O&M of turbines/generation, grid interconnection, and eventually nuclear/solar operating costs; economics unproven.
Capex
~$441M in Q1 2026; gross PP&E ~$1.43B. Funded via $785M of new equipment financing (incl. $500M from MUFG) and a $156M corporate financing commitment from Yorkville. First six Siemens SGT-800 gas turbines delivered (Port of Houston); site works underway (gas lines, water distribution, dual transmission interconnects).
Latest earnings
No consensus P&L to beat/miss; pre-revenue. Market focus is on liquidity runway and tenant signing, not EPS.
No revenue guidance. Management '90-day priorities' = secure a binding tenant agreement, manage liquidity, hire a permanent CEO, explore strategic partnerships with experienced data-center operators, and deliver first power. Lease revenue is contingent on tenant execution and not firmly dated.
- Net loss
- $189M ($0.30/sh)
- Cash (incl. restricted)
- $243M ($35.8M restricted)
- Total debt
- $421M
- Gross PP&E
- ~$1.43B
- Q1 capex
- $441M
- Power generation secured
- >2 GW (up to 17 GW targeted)
- Power permitted
- ~6 GW air permit obtained (+~5 GW filed)
- Equipment financing secured
- $785M (incl. $500M MUFG)
Growth drivers
- Landing a binding long-term anchor tenant (hyperscaler/neocloud) to unlock project financing and start lease revenue
- Bringing initial natural-gas generation online (>2 GW power generation secured near-term, up to 17 GW at full build-out)
- Permitting progress — ~6 GW TCEQ Clean Air Permit obtained, +~5 GW filed; NRC named the site an inaugural Environmental Impact Statement pilot for future nuclear
- Escalating AI-compute power demand outstripping grid interconnection queues, favoring behind-the-meter campuses
- Texas Tech University System partnership and Texas siting (land, permitting posture, workforce)
Bull & bear
A scarce, permitted, gigawatt-scale power+data campus that becomes strategically invaluable the moment one hyperscaler signs -- a call option on the AI power shortage trading well below its Oct-2025 debut.
- Power, not chips, is the binding constraint on AI; a shovel-ready site with ~6 GW permitted and dual interconnects is genuinely scarce and slow to replicate
- One binding anchor lease could unlock project financing, validate the REIT cash-flow model, and re-rate the equity violently higher off a depressed ~$5B cap
- Real assets on the ground: ~$1.43B gross PP&E, first Siemens turbines delivered, site works advanced -- not just a slide deck
- Deep-pocketed strategic acquirer (hyperscaler, IPP like Talen/Vistra/Constellation, or infra fund) could buy the campus outright; the ousted founder is publicly pushing a sale to maximize value
- Political and regulatory tailwinds (Perry, NRC pilot, Texas) lower some permitting friction versus peers
A pre-revenue, cash-burning developer with no signed tenant, a decapitated management team, active litigation, and a financing cliff -- where the entire thesis hinges on a hyperscaler deal that already fell through once.
- No revenue, no binding tenant, and a canceled ~$150M anchor arrangement (Dec 2025) with the March replacement target missed -- the core demand proof point is still absent
- ~$243M cash against ~$441M quarterly burn is a near-term funding cliff; survival depends on continuous external capital or a tenant, neither guaranteed
- Leadership vacuum: CEO removed and terminated for Cause, CFO resigned, no permanent CEO, plus a founder wrongful-termination suit and a board feud -- worst-possible backdrop for closing a multi-billion-dollar tenant
- Short-seller-flagged contingency triggers (land-lease revocation by 12-31-2026 absent a 200MW tenant, turbine resale after Nov 10 2026, air-permit legal challenge) plus fraudulent-transfer allegations
- Nuclear optionality is years and billions away and shouldn't be paid for today; even the gas plan carries heavy execution and interconnection risk
- Stock is down ~75% from its post-IPO peak for cause, not noise -- the market is pricing real going-concern/execution risk
What it is worth
Optionality / sum-of-the-parts on a pre-revenue asset base; no earnings, FCF, or contracted revenue to anchor a multiple.
No binding tenant before contingency deadlines -> land-lease revocation (Texas Tech, 12-31-2026), turbine resale, permit challenges, and a financing cliff drive the equity toward distressed/asset-recovery value well below the current cap.
Muddle-through: partial power online and non-binding LOIs, continued dilution/financing to bridge to a first tenant; stock stays volatile and range-bound around a low-single-digit-billions to ~$5B cap pending a signed deal.
Binding hyperscaler anchor lease + project financing (or an outright strategic/hyperscaler acquisition of the campus) revalidates the REIT cash-flow model and re-rates the equity materially above current levels toward its debut valuation range.
At ~$5.1B market cap the equity is being valued as a call option on landing a hyperscaler tenant and executing a multi-GW build, well above the ~$1.43B gross PP&E and ~$243M cash on the balance sheet. With no signed revenue, thin liquidity vs. burn, and governance/litigation overhang, valuation is speculative and binary: it re-rates hard on a binding anchor lease or strategic sale, and de-rates toward asset/liquidation value if the tenant and financing don't materialize. Dual Nasdaq+LSE listing; figures in USD.
SWOT
Strengths
- Large, permitted, single-site campus in Amarillo, TX with dual transmission interconnects and ~6 GW air permit already in hand -- hard to replicate quickly
- Vertically integrated 'private grid' thesis directly targets the #1 constraint on AI build-outs (power + interconnection speed)
- Political/regulatory heft — co-founder Rick Perry (former U.S. Energy Secretary), NRC EIS pilot selection, Texas siting
- Capital access despite turmoil — $785M gross IPO incl. greenshoe (~$745.6M net) at $21/sh in Oct 2025, plus $785M equipment financing (incl. $500M MUFG) and a $156M Yorkville facility
Weaknesses
- Zero revenue and no binding anchor tenant — a prior ~$150M tenant arrangement was terminated Dec 2025 and no replacement met the March target
- Severe governance instability — CEO Toby Neugebauer removed Apr 17 and terminated 'for Cause' Apr 30; CFO Miles Everson resigned Apr 19 in the 'Fermi 2.0' reset; no permanent CEO; wrongful-termination litigation and a board-vs-founder feud
- Thin liquidity (~$243M cash) against ~$441M/quarter burn — a self-evident funding cliff without near-term financing or a tenant
- Heavy share-based comp / dilution optics and short-seller allegations of fraudulent transfers; large share-block overhang since IPO
Opportunities
- Structural AI power shortage — hyperscalers and neoclouds actively seeking gigawatt-scale behind-the-meter power they cannot get from congested grids
- Optionality to add nuclear (SMR/large) and solar atop gas, potentially commanding premium 'clean firm' power contracts
- Possible strategic partnership, JV, or outright sale of the campus to a deeper-pocketed operator (ousted founder and board both publicly pushing value-maximization, though they disagree on a sale)
- First-mover 'energy campus REIT' category could re-rate sharply on a single credible hyperscaler signing
Threats
- Tenant-contingency cliff — short-seller Fuzzy Panda flagged that Texas Tech can revoke the land lease if a 200MW tenant isn't signed by 12-31-2026, turbines become saleable after Nov 10 2026 if uncommitted, and the TCEQ air permit faces a legal challenge
- Execution/timeline risk on multi-GW gas and (especially) nuclear — capital, supply chain, and NRC timelines measured in years
- Rising rates / risk-off could shut off the external capital the model depends on before revenue
- AI-capex/power-demand cyclicality — if hyperscalers slow behind-the-meter deals or build their own, demand thesis weakens
- Reputational and legal overhang (short-seller fraud allegations, founder litigation) raising cost of capital
Moats, dependencies & bottlenecks
Moats
~6 GW TCEQ air permit, dual transmission interconnects, and a single large Texas parcel are hard to assemble quickly -- but the moat is a land/permit head-start, not a durable technology or contract advantage, and the air permit faces legal challenge and the ground lease is tenant-contingent.
Rick Perry pedigree and NRC EIS-pilot selection ease some friction; not a cash-flow moat and fades if execution stalls.
Weak-to-moderate Owning generation + campus could lock in tenants, but only once contracts exist; today it is a thesis, not a realized switching cost. IPPs (Talen/Vistra/Constellation) can replicate the power side with existing nuclear fleets.
Dependencies
Commercial / existential No revenue, financing, or arguably survival without one; a prior ~$150M arrangement was terminated Dec 2025. Texas Tech can revoke the land lease if a 200MW tenant isn't signed by 12-31-2026, and turbines become saleable if uncommitted.
Burn far exceeds cash on hand; company relies on equity, equipment debt (MUFG $500M), and facilities (Yorkville $156M). Any risk-off shift or reputational overhang raises cost/availability of capital.
SGT-800 gas turbines are the near-term generation backbone; delivery/commissioning timeline gates first power. Turbines reportedly saleable after Nov 10 2026 if uncommitted.
Air permits (~6 GW issued, ~5 GW filed) face a TCEQ legal challenge, NRC nuclear licensing is years out, and interconnection approvals all gate the build.
No permanent CEO; CFO resigned; founder litigation and board feud impair deal-closing and financing credibility.
Advantages
- Head-start on a permitted, interconnected, gigawatt-scale Texas campus few competitors can match today
- Direct exposure to the AI power bottleneck -- the scarcest input in the AI build-out
- Political/regulatory access (Perry, NRC pilot) and a Texas Tech University System partnership
- Already-deployed real assets (~$1.43B PP&E, first turbines in-country) rather than a paper project
Weaknesses
- Pre-revenue with no signed tenant and a canceled anchor deal
- Near-term funding cliff and heavy dilution/burn
- Leadership decapitation plus active litigation and short-seller fraud allegations
- Business model unproven end-to-end; power economics and lease pricing untested
- Nuclear optionality is distant, capital-intensive, and uncertain
Bottlenecks
- Securing the first binding anchor tenant — everything (financing, land-lease retention, timeline) cascades from it
- Liquidity runway vs. ~$441M/quarter burn
- Filling the permanent CEO/CFO vacuum to restore counterparty and investor confidence
- Turbine delivery/commissioning and grid interconnection to deliver first commercial power
- Multi-year, multi-billion-dollar path (and NRC licensing) for the nuclear component
Top signals & trends
Top signals
Removed the core demand validation; no binding tenant since.
Governance instability; no permanent CEO; founder wrongful-termination litigation and board feud follow.
Alleged fraudulent transfers and flagged land-lease revocation (200MW tenant by 12-31-2026), turbine resale, and air-permit legal risks; stock fell sharply.
Mildly bullish · Shows continued lender willingness to fund hard assets despite turmoil.
Tangible permitting progress that de-risks part of the build, though the air permit faces a legal challenge.
Prices in heavy risk; also sets up a violent re-rate on a single tenant signing (high-variance).
Trends
Core tailwind; hyperscalers increasingly pursue behind-the-meter gigawatt-scale power -- exactly Fermi's pitch.
Validates the demand thesis but shows incumbents with existing generation are landing the deals first.
Near-term gas generation is the pragmatic path to first power ahead of nuclear; Fermi is squarely on it.
Punishes pre-revenue, capital-hungry developers; raises Fermi's cost of survival capital.
Positive (long-dated) · Supports the eventual nuclear leg but on a multi-year horizon.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
SGT-800 gas turbines -- the near-term generation backbone; delivery timeline gates first power.
$500M of the $785M equipment financing; key debt provider.
$156M corporate financing commitment (equity/convertible facility).
Leading gas-turbine / grid-equipment supplier and potential vendor for scaling generation.
Prospective hyperscaler tenant (Azure); none signed. Already contracting nuclear elsewhere (Constellation).
Prospective tenant; already doing behind-the-meter nuclear with Talen -- shows appetite for Fermi's model.
Prospective hyperscaler tenant; active in nuclear/SMR (Oklo/Kairos) power deals.
Prospective large-scale AI power/campus tenant.
Prospective tenant; aggressive AI-datacenter (Stargate) capacity buyer.
Neocloud actively leasing gigawatt-scale AI capacity -- archetype of Fermi's target tenant.
Neocloud expanding US AI capacity; another prospective tenant.
Nuclear IPP already powering Amazon data centers behind-the-meter; a proven, cash-generative version of Fermi's thesis.
Largest US nuclear fleet; signed Microsoft PPA (Three Mile Island restart). Incumbent clean-firm power for AI.
Large gas+nuclear generator courting data-center load; can deliver power without building a campus from scratch.
Closest listed analog -- builds large AI/HPC data-center campuses with power procurement; already signing hyperscaler leases (e.g., CoreWeave).
Advanced-nuclear developer targeting data-center power; competes for the future nuclear-powered-AI narrative and capital.
SMR vendor; competes for the SMR-for-data-centers thesis and mindshare.
On-site fuel-cell power for data centers -- an alternative behind-the-meter power path with faster deployment.
Incumbent data-center REIT with scale, tenants, and power procurement muscle; the established version of the REIT model.
Global data-center REIT with deep hyperscaler relationships and power procurement; established competitor for AI capacity.
Private; builds power-dense AI data centers (Stargate Abilene, TX) -- direct behind-the-meter competitor for hyperscaler campuses.