Fermi Inc. filed its first 10-K annual report, confirming what the committee's 6-lens analysis anticipated: massive physical infrastructure progress colliding with a complete absence of binding commercial commitments. The company has deployed $935M in construction-in-progress, secured a 6 GW air permit, confirmed the AP1000 Westinghouse reactor design, and closed $500M in MUFG financing. But the only financial commitment from a tenant — a $150M AIAC agreement — was terminated in December 2025. Revenue remains $0. The stock fell 24% on earnings day to ~$5.12, extending the decline to 76% from its $21 IPO price.
The Core Tension: Infrastructure Without Tenants
Fermi's FY2025 earnings present a genuine paradox. The infrastructure progress is among the most tangible of any pre-revenue energy company: 42 Siemens SGT-800 gas turbines secured (with 5 installed), a 6 GW air permit finalized in February 2026 with an additional 5 GW filed, and the first NRC large reactor application in 15+ years with 4 AP1000 Westinghouse reactors confirmed. The MUFG $500M facility closed post-year-end, fully refinancing the Macquarie term loan and eliminating the 2026 maturity wall that was the single largest near-term risk in our prior analysis.
But the commercial validation gap has widened, not narrowed. The $150M AIAC (tenant advance for infrastructure construction) was terminated in December 2025 — the only binding financial commitment from any tenant. CEO Neugebauer characterized prospects as being in a “contracting phase” and said the air permit was what “really turned our shoppers into buyers.” But there are no buyers yet. Zero binding leases. Zero signed commitments. The board is explicitly withholding tenant timeline guidance.
Signal Change: 1 Reclassified, Multiple Confidence Upgrades
Our 6-lens update produced one signal reclassification out of seven monitored signals. The remaining six confirmed at prior levels with confidence upgrades from MEDIUM to HIGH based on the substantially richer evidence base from the first 10-K:
The EXPECTATIONS_PRICED reclassification from EXUBERANT to STRETCHED reflects the 76% stock decline compressing market cap from ~$13B to ~$3.2B. At $5.12, Fermi trades at 2.3x book value and 1.8x total capital assembled — the market no longer implies near-certainty of the full 17 GW vision. However, $3.2B for a pre-revenue company with zero binding tenants and a terminated AIAC still requires meaningful execution milestones to justify. Notably, 8 of 8 sell-side analysts maintain Buy ratings at an average $29 target (466% upside) — a disconnect that itself functions as a Myth Meter signal.
Key Developments: De-Risking vs. Commercial Void
| Development | Impact | Assessment |
|---|---|---|
| MUFG $500M closed | Refinanced Macquarie, eliminated 2026 maturity wall | Positive |
| 6 GW air permit finalized | Gas-side regulatory bottleneck removed | Positive |
| AP1000 reactor design confirmed | NRC COLA filed for 4 reactors, pilot program selected | Positive |
| NRC environmental review initiated | Applicant-prepared EIS may reduce timeline | Positive |
| $150M AIAC terminated | Only binding tenant commitment eliminated | Negative |
| Zero binding tenant leases | Commercial validation gap widened | Negative |
| Lock-up expiry (March 30, 2026) | Pre-IPO investors free to sell | Risk factor |
| REIT 5/50: Neugebauer at 38% | Must sell down, hired advisor for block sale | Structural |
CEO Neugebauer described a strategic shift toward reverse due diligence — Fermi is now evaluating tenant MEP (mechanical, electrical, plumbing) readiness rather than simply seeking tenants. The company hired an ex-Meta executive specifically for this purpose. Management targets 500+ MW minimum deals with 2–3 tenants for the first 2 GW. Modular MEP via factory-built solutions (Schlumberger visit noted) may compress deployment timelines. There is also “significant engagement by chip makers” — a new tenant category beyond hyperscalers.
Reading the Numbers: $486M Loss in Context
The headline $486M net loss is primarily non-cash. A $104M Q4 SBC surge is explained by IPO-triggered vesting acceleration, not ongoing compensation. The charitable donation (previously flagged in our analysis) remains opaque. The true operating cash burn is ~$34M/year, which against $409M in cash provides roughly 4 years of runway at current rates — though capital commitments for Phase 0 and Phase 1 development require substantially more than available cash.
The balance sheet tells the asset story more clearly than the income statement: $1.4B total assets anchored by $935M in CIP with $1.1B in equity. The ~628M shares outstanding at ~$5.12 produces the ~$3.2B market cap. Management saved $27–30M by clearing SGT-800 turbines through customs before the Supreme Court tariff ruling — a genuine operational execution win.
Prediction Markets: 3 Resolved
Three prediction markets resolved with this earnings report at an average Brier score of 0.1432 — solid calibration with one excellent call, one strong call, and one moderate call:
The CIP impairment call was the ensemble's best (Brier 0.0144) — the models correctly assessed that management would not impair assets during an active construction program with increasing capital commitments. The MUFG closing prediction (Brier 0.1849) was solid: the ensemble assigned moderate probability and the deal closed post-year-end. The reactor design disclosure (Brier 0.2304) was directionally correct at 52% but underweighted the regulatory filing dynamics that made disclosure near-certain once the NRC COLA was submitted.
Today: Lock-Up Expiry and Structural Pressures
The pre-IPO investor lock-up expired today, March 30, 2026. This coincides with two structural selling pressures: the Neugebauer family holds 38% of shares and must sell down for REIT 5/50 compliance (no single holder may own >50%, and top 5 holders combined must exceed 50%). They have hired an advisor for an “accretive” block sale, though the mechanism for making a mandatory sell-down accretive is unclear.
Additionally, the Texas Tech ground lease imposes a 200 MW tenant + Notice to Proceed deadline by end of 2026 — only 9 months away. Failure to meet this condition could jeopardize the land rights underlying the entire Project Matador development.
What to Watch Next
Sources and methodology
Primary sources: FRMI 10-K Annual Report (FY2025), filed with SEC on March 25, 2026; Fermi Inc. FY2025 Earnings Call Transcript (March 26, 2026); NRC COLA Application and Notice of Intent (March 20, 2026).
Analysis methodology: Multi-lens committee analysis using the Runchey Research AI Ensemble. Six lenses (Fugazi Filter, Stress Scanner, Moat Mapper, Regulatory Reader, Myth Meter, Black Swan Beacon) updated with earnings data. Each lens employs structured discourse between multiple AI models with persona-based orchestration.
Prediction markets: 9-model ensemble generating binary probability forecasts scored via Brier metric (0.0 = perfect, 1.0 = worst).
Full 6-lens analysis with signal table, cross-lens reinforcements, monitoring triggers, and committee discourse
FRMI Full Analysis