Why Fermi Stock Collapsed 41% Last Month

Shares fell 41% in March and are down more than 80% from their highs. Fermi has zero revenue, burned roughly $600M in free cash flow in its first year, lost an initial tenant that pulled a ~$150M upfront commitment, and faces massive capex to build Project Matador (targeting 17 GW) after raising $750M in its IPO; upcoming lock-up expirations add further downside risk.

Analysis

A boutique REIT model that sells large, bespoke infrastructure (power plants and private grids) into hyperscale AI demand is misaligned with how those customers procure risk today: they prefer rapid, contracted delivery and minimal development execution risk. That mismatch creates persistent offtake and financing friction — hyperscalers will pay a premium for near-term availability but are unlikely to commit long-term capital to a single unproven developer without creditworthy guarantees or material pricing concession.

Second-order winners will be incumbent transmission owners, modular turbine suppliers and short-term power providers who can deliver capacity on 6–24 month timeframes; they become natural strategic partners or acquirers rather than direct competitors. Conversely, capital structures that compress early revenue optionality (REIT taxonomy, high fixed-cost buildouts) will see amplified equity volatility and refinancing risk if even one marquee counterparty delays.

Key catalysts to watch are threefold and operate on different horizons: near-term (weeks–months) — financing commitments or strategic JV announcements from investment-grade counterparties that would de-risk cash flow visibility; medium-term (6–24 months) — signed offtake/take-or-pay contracts and construction milestones that materially derisk execution; long-term (24+ months) — regulatory support (grants/PTCs/loan guarantees) or technology breakthroughs that lower capex/unit. Tail risks include counterparty pull-through failure, sustained capex inflation, and higher-for-longer rates that blow out IRRs and force equity dilution.

From a portfolio perspective, the simplest efficient response is to separate exposure to AI demand (structural long) from exposure to execution risk (short); that preserves upside from compute growth while isolating idiosyncratic project execution and capital-structure risk. Monitor three datapoints as trade triggers: a) a funded take-or-pay from an investment-grade hyperscaler, b) a material utility JV or offtake backstop, c) a suspected insider sale wave around lock-up expiries — any of which should materially reprice equity risk premia.

Key Decisions for Investors

• Short FRMI equity (2–6 month trade), position size 1–2% NAV. Use a tactical entry after any short-lived bounce; set stop-loss at 25% adverse and target a 30–60% downside to capture execution/lock-up and offtake risk. Rationale: limited near-term cash flows with concentrated execution risk — payoff asymmetric if no new secured contracts emerge.
• Pair trade: Long NVDA (leveraged exposure via 12–18 month LEAP calls) / Short FRMI (equity) in equal notional terms. This isolates secular AI upside while hedging project execution and capital-structure risk. Risk: NVDA idiosyncratic drawdowns; cap position size so pair correlation break does not exceed 4% NAV loss.
• Long regulated transmission/utilities (e.g., NEE, AEP) — 12–36 month hold. Buy selective equities or IG corporate bonds sized 2–4% NAV. Expected return 15–35% if grid upgrade cycle accelerates; risk is demand softness or regulatory delays, hedge with 1–2% NAV protection (puts).
• Buy FRMI 3–6 month put spread (cost-limited options trade): buy a nearer-OTM put and sell a deeper-OTM put to cap premium. Allocate notional ≈0.5–1% NAV; payoff multiplies if sentiment and funding momentum fail. This expresses short conviction with defined max loss.