Fermi Inc. Common Stock

Executive Summary

HOLD, Low conviction. Rated HOLD, Low conviction. FRMI is no longer a pure concept stock after signing a binding TensorWave agreement for up to 650 MW and outlining first power on July 1, 2027, but the equity still sits in a funding-and-execution gap with negative FY2025 EPS of -$1.13, FCF of -$0.4B, and no formal company-wide EBITDA or CapEx guidance. At $5.71, versus the engine base value of $6.03, upside is only 5.6%, which is not enough to pay for the very high project-delivery risk implied by the stock's -84.6% drawdown from the high and repeated earnings misses.

Investment Thesis

FRMI has crossed the line from story stock to partially validated project, and that matters. The binding TensorWave agreement for up to 650 MW, the disclosed $6.5B of initial-term revenue on Phase 1, and a specific first-power date of July 1, 2027 are concrete milestones that did not exist in the same form before. The Hillcore alliance also matters because it expands optionality by 2.6 GW under a BOOT structure with no upfront FRMI capital outlay. Those are not cosmetic developments. They are the first pieces of evidence that FRMI can monetize AI-linked power scarcity.

The problem is that the equity still has to bridge almost all of the hard parts. FY2025 revenue remains listed at $0.0B, EPS was -$1.13, Q2 2026 EBITDA was -$0.03B, and FY2025 FCF was -$0.4B. Investors still do not have formal company-wide guidance on EBITDA, margins, or total project CapEx. That makes underwriting fair value difficult, especially when the stock already trades in a market that is rewarding delivered AI infrastructure, not just promised infrastructure. FRMI may well become valuable, but the current valuation case rests more on milestone confidence than on forecastable earnings power.

Management execution is the fulcrum. On the positive side, management said, "To date, we've delivered on all 5 commitments, including the one that matters most," and the quarter did include visible operational progress, including turbine deliveries and a signed anchor customer. On the negative side, the earnings record is rough. FRMI missed consensus by -3569.7% in November 2025, -2526.8% in March 2026, and -784.2% in May 2026, before beating by 12.0% in August 2026. That is not the profile of a management team that has earned the benefit of the doubt on a capital-intensive 2027 schedule.

At $5.71, the stock is also no longer being judged in a vacuum. It sits 84.6% below the $36.99 high and 37.1% below the 200-day average of $9.07, which shows how much prior optimism has already been broken. That collapse creates upside if financing closes and 2027 milestones hold, but it also signals that investors have learned to demand proof. The case depends on three things: closing the backstop and project finance, preserving the 2027 delivery schedule, and converting one anchor contract into a broader booked pipeline. Until those pieces are firmer, FRMI looks more like a watchlist name than an aggressive buy.

Key Metrics

Financial Strength

FRMI's balance sheet reads better on historical leverage than on self-funding capacity. Reported debt is $0.0B and debt-to-equity is 0.00, but that understates economic funding risk because the company is still pre-scale and burning cash, with negative FCF of $0.4B and negative Q2 EBITDA of $0.03B. The upsized convert improves runway and buys time for negotiations, but until project finance is secured and revenue begins in late 2027, financial strength depends more on access to capital than on internally generated cash. Margin trajectory cannot yet be judged because management has not provided formal EBITDA or CapEx guidance and reported revenue is still not meaningful.

Competitive Position

The current moat is access-based, not earnings-based. FRMI's advantage is control over scarce power capacity, a signed anchor customer, long-duration contract structure, and a pass-through on fuel and variable energy costs, which reduces commodity risk once the assets are operating. Durability remains unproven because a single signed customer and a pre-revenue model do not yet constitute a fully defended franchise, so the moat becomes real only if the company delivers 210 MW on time in 2027 and turns the 4.8 GW pipeline into repeat business.

Management & Guidance

Management's latest guidance is specific on physical delivery and customer timing but thin on economics. The company guided to 222 MW of Phase 1 gross power with a 15-year initial term, 210 MW first power on July 1, 2027, 640 MW nameplate capacity by Q4 2027, and phased revenue commencement in late Q3/early Q4 2027. That operating specificity is a step forward, and the quarter did include delivery on all 5 self-set 90-day objectives, but credibility is only partial because FRMI still has no formal company-wide EBITDA, margin, or total project CapEx guide and its earnings beat-miss history is poor. The result is a management team that has improved commercial execution but has not yet earned full financial credibility.

Segment Analysis

• Power delivery contracts (Revenue not yet meaningful): The first disclosed commercial block is 222 MW of Phase 1 gross power tied to $6.5B of initial-term revenue over 15 years, with first revenue expected in late Q3/early Q4 2027.
• TensorWave customer program (Up to 650 MW contracted capacity across 3 phases): This is the anchor demand relationship and currently the clearest proof that FRMI's AI-focused power capacity has a buyer.
• Hillcore alliance / campus expansion (2.6 GW incremental under BOOT structure): Adds large optional capacity with no upfront FRMI capital outlay, expanding total campus capacity to 4.8 GW.
• Geography (Concentrated in Texas): The thesis is tied to Texas power availability, project execution, and data-center demand around the Houston campus.

Capital Allocation

Capital allocation is still about survival and project enablement rather than shareholder return. FRMI generated negative FCF of $0.4B, spent $0.4B on capex, and paid $0.0B in buybacks and $0.0B in dividends, which is appropriate for a build stage asset but leaves no margin for execution mistakes. The $431M convertible, with approximately $382M of net proceeds after the capped call, was a pragmatic move to extend runway, though it also confirms dependence on external capital until project cash flows arrive.

Positioning & Flows

Ownership is mixed but notable for alignment, with insiders holding 31.0% and the top 3 institutions owning 17.3%, which supports commitment but also means public float sentiment can still swing hard on milestone news. Short interest is elevated at 47.1M shares, or 7.81% of shares with 2.4 days-to-cover, showing a meaningful skeptical camp after the stock's collapse. Options positioning is modestly supportive near term, with positive GEX and zero-gamma at $4.98, about 12.4% below spot, which may dampen immediate downside volatility but does not change the fundamental financing and delivery debate.

Bull Case

• Commercial risk dropped materially after management signed a binding TensorWave agreement for up to 650 MW across 3 phases, the first concrete demand signal behind the Texas AI power buildout.
• Contract economics are now visible at least at Phase 1, with $6.5B of initial-term revenue tied to 222 MW of gross power under a 15-year initial term.
• The operating timeline tightened meaningfully, with management targeting 210 MW of first power on July 1, 2027 and 640 MW delivered by Q4 2027, while three Siemens F-class turbines have already arrived in Houston.
• Structure improved with the Hillcore alliance adding 2.6 GW under a BOOT model with no upfront FRMI capital outlay, and fuel or variable energy costs are fully passed through to customers.

Bear Case

• The business still lacks current earnings support, with FY2025 EPS of -$1.13, FCF of -$0.4B, and Q2 2026 EBITDA of -$0.03B.
• Execution credibility is weak because FRMI missed Street EPS by -3569.7% twice in November 2025, -2526.8% in March 2026, and -784.2% in May 2026 before a modest beat in August.
• Management has not provided formal company-wide guidance for EBITDA, margin, or total project CapEx, leaving investors unable to underwrite 2027 to 2028 earnings power with confidence.
• The stock setup is fragile, down 84.6% from the $36.99 high and still carrying 7.81% reported short interest, reflecting skepticism that financing and power-delivery milestones will land on schedule.

Valuation & Price Target

Engine price target $6.03 (+5.6% vs current), Low confidence.

• Quality-adjusted: $6.03 (20% weight)

The rating stays aligned with the quant baseline of HOLD because the engine's $6.03 target implies only 5.6% upside from $5.71, and the underlying evidence still lacks the profitability visibility needed to press harder. There is a real case to diverge upward based on the 650 MW binding deal, $6.5B initial-term revenue, and improved liquidity from the $431M convert, but that is offset by FY2025 EPS of -$1.13, FCF of -$0.4B, no formal EBITDA or CapEx guidance, and an ugly miss record before the latest quarter. Sell-side consensus sits at Strong Buy, but that is low-signal context here and not enough to outweigh the combination of thin valuation upside and binary 2027 execution risk.

Risk Assessment

Risk score 44/100 (Moderate). The biggest risk is that project finance and the expected backstop fail to close on time, which would threaten the planned 210 MW first power date on July 1, 2027 and undermine the value of the $6.5B initial-term contract.

The Bottom Line

FRMI rates HOLD at Low conviction. The company finally has proof that its AI power concept can attract a real customer, and the $6.5B contract plus 2027 milestones create legitimate upside if financing and construction stay on track. That said, the stock is still a project-delivery security with weak current fundamentals, missing long-term financial guidance, and only 5.6% upside to the engine base case of $6.03. A move to BUY would require firmer evidence that the backstop and project finance are closed and that the July 1, 2027 first-power date remains fully funded and executable. A slip in either would push the view toward SELL.