Fermi: Hyper High Risk, Hyper High Reward

Fermi is being framed as a highly speculative AI infrastructure play, with a 17 GW Texas power campus and a current valuation of 16x forward P/FFO. The upside case hinges on flawless execution, rapid ramp-up, and landing major AI customers such as OpenAI, with projected 2028 revenue of $4.5B and FFO of $2.0B-$2.5B implying a possible 2x forward P/FFO. Financing, execution, and client acquisition risks remain acute, tempering the investment case.

Analysis

FRMI is less a real-estate story than a long-duration project-finance option on AI power scarcity. The market is likely underpricing the convexity if management can lock anchor tenants early, because the value inflection is not linear with occupancy: once one credible hyperscaler signs, financing costs, permitting confidence, and customer de-risking all improve together. That creates a winner-take-most dynamic where early client wins could compress the implied cost of capital by several hundred bps and re-rate the equity far faster than the physical buildout.

The second-order beneficiaries are the picks-and-shovels stack: grid interconnect specialists, gas turbine/backup power suppliers, electrical switchgear, cooling, and regional transmission contractors. Conversely, pure-play AI infra names without land/power control are vulnerable if FRMI proves it can package power, acreage, and permitting into a single financed product; the scarcity premium migrates from chips to electrons and time-to-energize. The real competitive moat is not acreage but entitlement speed and utility relationships, which are difficult to replicate and can make adjacent sites effectively stranded even if they are cheaper on paper.

The main tail risk is funding dilution before the project becomes financeable at scale. If client signing slips by 6–12 months, the equity can reprice sharply lower because the bull case depends on a narrow timing window where capex precedes cash flow by years; that gap will likely force repeated capital raises or higher-cost debt. Another underappreciated risk is that AI demand may not want single-tenant mega-campuses if model training shifts toward distributed, shorter-cycle inference loads, which would reduce the premium for one giant campus and raise lease-up risk.

The contrarian view is that the market may be too focused on headline upside and not enough on the embedded option value from scarcity, especially in a regime where power availability is the binding constraint. If management can demonstrate a bankable anchor tenant and credible phased financing, the stock could rerate well before first revenue, because public markets typically pay up for de-risked capacity under construction. But absent that proof, this remains a high-beta financing trade rather than a fundamentals trade, and the gap between story and monetization is where the downside sits.

Key Decisions for Investors

• Do not buy FRMI outright ahead of a signed anchor-tenant announcement; use it as a catalyst trade only. Best entry is after a financing package plus first material customer disclosure, when the probability-weighted dilution risk falls materially.
• For aggressive upside exposure, buy medium-dated call spreads on FRMI rather than common stock to cap financing/dilution risk. A 6–12 month structure offers favorable convexity if the market re-rates on client wins, while limiting loss if lease-up slips.
• Pair trade: long FRMI / short a public AI infra developer that lacks secured power or entitlements. This expresses the scarcity premium and isolates the differentiator as permitting and grid access rather than generic AI demand.
• If FRMI rallies on narrative alone without signed customers, fade strength into the move. The stock should be treated as a multiple-expansion trade only after execution milestones, not before; upside is large but path dependence is extreme.
• Monitor announcements from hyperscalers and model labs over the next 1–2 quarters for any Texas power commitments. If a major client signs nearby, re-underwrite FRMI immediately because the probability of follow-on capacity demand and lower financing cost rises sharply.