Fermi is a pre-revenue, behind-the-meter “HyperGrid” REIT attempting to sell guaranteed gigawatt-scale AI power—where speed-to-power is the moat, and execution is the entire thesis.
Fermi Inc., operating commercially as Fermi America, is a development-stage infrastructure enterprise purpose-built to resolve the primary bottleneck of the generative artificial intelligence era: the availability of high-density, gigawatt-scale power.[1, 2, 3] Founded in January 2025 and headquartered in Amarillo, Texas, the company is developing Project Matador, a multi-gigawatt energy and data center campus that integrates private power generation with high-performance computing (HPC) facilities.[2, 3, 4] Unlike traditional data center providers that rely on local utilities for power, Fermi is pioneering the "HyperGrid" concept—a private, behind-the-meter electric network designed to provide highly redundant, off-grid power to hyperscale tenants.[3, 5]
The company generates its economic value through a Real Estate Investment Trust (REIT) structure, aiming to offer investors tax-efficient exposure to the massive capital expenditures occurring in the AI infrastructure sector.[1, 6, 7] Fermi’s revenue model is centered on long-term, triple-net (NNN) leases where rental income is increasingly decoupled from traditional square-footage metrics and instead tied to reserved power capacity and megawatt-hour utilization.[6] This "power-linked rent" approach aligns the company’s revenue with the operational needs of hyperscalers—such as Microsoft, Google, and Amazon—and specialized "neocloud" providers like Fluidstack or CoreWeave, who require vast amounts of electricity for large language model (LLM) training and inference.[7, 8, 9]
Fermi’s core products consist of "AI-ready powered shells"—massive industrial buildings equipped with the specialized cooling and power distribution required for modern GPUs—and direct access to an integrated energy supply.[3, 6] This energy supply is designed to be highly diversified, incorporating combined-cycle natural gas, solar generation, battery energy storage, and a long-term roadmap for advanced nuclear power through Westinghouse AP1000 reactors and Small Modular Reactors (SMRs).[3, 10, 11] The primary end markets for these services are concentrated in the United States, specifically within the defense, national security, and commercial AI sectors, where energy independence and high availability are paramount.[3]
Customers choose Fermi over established competitors primarily due to "speed-to-power".[7] Traditional data center hubs, such as Northern Virginia or Santa Clara, are currently grappling with grid congestion that can lead to 5-to-7-year wait times for new power interconnections.[12, 13] By owning its own generation and operating behind-the-meter in the energy-rich Texas Panhandle, Fermi seeks to bypass these bottlenecks, offering hyperscalers the ability to deploy compute clusters at a scale and velocity that traditional utility-bound providers cannot match.[7, 10] As of early 2026, the company is in a transition phase dubbed "Fermi 2.0," which involves a strategic leadership overhaul aimed at moving from a project-development startup to a scaled infrastructure operator.[14, 15]
The strategic underpinning of Fermi Inc. is the recognition that the "physics of AI" has outpaced the "regulation of the grid." The company’s growth is driven by the divergence between the exponential demand for compute and the linear, regulated expansion of traditional power utilities.
Fermi is not merely a real estate company; it is an integrated energy-computing utility. The primary product sold to investors and tenants is "guaranteed capacity." In a traditional lease, an investor analyzes the price per square foot; however, for Fermi, the relevant metric is the price per kilowatt per month ($/kW/month).[6, 12] This model reflects the reality that in an AI data center, the cost of the building shell is a fraction of the cost of the power systems and the IT equipment housed within it.[13, 16]
The "AI-ready powered shells" at Project Matador are designed to support rack densities exceeding 100 kW, which is the threshold where traditional air-cooling systems fail.[17, 18, 19] Fermi provides the infrastructure for liquid cooling and direct-to-chip heat rejection, which can reduce facility energy use by up to 40%.[18] Furthermore, the company offers "grid-independent" energy, meaning that even in the event of a regional grid failure, the on-site natural gas turbines and battery systems can maintain uptime for critical AI workloads.[3, 5] This is particularly attractive for "training" workloads, where a sudden power loss can result in weeks of lost progress and millions of dollars in wasted compute time.
Fermi’s competitive moat is constructed from a combination of regulatory, geographic, and structural barriers.
The Total Addressable Market (TAM) for Fermi is defined by the global "infrastructure supercycle" driven by AI. The total AI infrastructure market—including hardware, storage, and power—was valued at $101.17 billion in 2026 and is projected to reach $202.48 billion by 2031, representing a 14.89% CAGR.[18] More specifically, the data center sector is expected to require $3 trillion in investment by 2030, with roughly 100 GW of new capacity needed globally to satisfy demand.[13]
Project Matador’s 11 GW capacity (potentially expanding to 17 GW) positions Fermi to capture a significant share of the North American wholesale data center market.[11, 20, 21] In the U.S. alone, data center capacity is projected to increase by 200 GW through 2032.[16] If Fermi captures just 5% of this new capacity, it would represent a gigawatt-scale footprint that dwarfs most established REITs. The current supply-demand imbalance in the power market suggests that pricing power will remain with developers like Fermi who have "speed-to-power" capabilities.[9, 13]
Fermi is positioned at the intersection of several established industries, creating a unique competitive profile.
Fermi appears to be "holding ground" in terms of its permitting and physical construction milestones but has recently lost ground in terms of market valuation and leadership stability due to the "Fermi 2.0" restructuring and the exit of its anchor tenant.[28, 29, 30, 31]
Analyzing Fermi Inc. requires a shift from traditional P&L metrics to a "capital formation and deployment" framework, as the company is currently in a pre-revenue construction phase.[21, 28]
Fermi announced its fiscal year 2025 results on March 30, 2026.[2, 32, 33] The report covered the period from its inception on January 10, 2025, through December 31, 2025.[21, 28]
The latest earnings materials revealed a meaningful shift in the development timeline. Management previously targeted 1.1 GW of power online by the end of 2026. In the March 30, 2026, report, this was revised to a target of 2.0 GW of gas-fired and grid-supplied power by the end of 2027.[6] This delay is attributed to the need for "tenant MEP (Mechanical, Electrical, and Plumbing) readiness" and adjustments in the procurement schedule for large-scale turbines.[6, 36]
In his final shareholder letter before his departure, co-founder and former CEO Toby Neugebauer noted that "Project Matador is no longer a concept, but a $1.4 billion physical reality," pointing to the 4.2 miles of gas line and 11.3 miles of fencing already installed.[2, 21, 36] However, the absence of a "memorialized" binding lease agreement during the fourth-quarter call was a major point of concern for analysts, leading to a sharp 20% decline in the stock price on the day of the announcement.[4, 28, 37]
A more significant impact on the stock occurred on April 17, 2026, when the company announced the "Fermi 2.0" strategic evolution.[14, 15] This involved:
* The departure of co-founder and CEO Toby Neugebauer.[14, 30, 38]
* The resignation of CFO Miles Everson, who transitioned to a board role.[14, 15]
* The appointment of Marius Haas (former Dell executive) as Chairman.[14, 15]
* The creation of an "Interim Office of the CEO" led by Jacobo Ortiz and Anna Bofa.[14, 38, 39]
Following this news, shares fell another 20% to 30% in after-hours trading, reflecting intense investor unease regarding management continuity and the underlying reason for the abrupt exits.[30, 31, 37] Analyst reactions were split: Stifel maintained a "Buy" rating and a $29 target, citing the company's strong cash position and the quality of the new board.[40] Conversely, UBS slashed its price target from $30 to $8, highlighting the execution risk associated with the leadership void and the lack of tenant traction.[29, 37, 41]
Traditional valuation multiples like P/E are not applicable. Instead, investors focus on Enterprise Value (EV) relative to "Future Capacity" and Net Tangible Book Value (NTBV).[42, 43]
| Metric | Current Value / Estimate |
|---|---|
| Market Capitalization | ~$3.28 Billion [21, 35, 37] |
| Enterprise Value (EV) | ~$3.0 Billion [28, 44, 45] |
| EV / Capacity (Phase 1 Target) | $1.5 Million per MW (Base on 2.0 GW target) |
| Pro Forma NTBV | ~$1.53 per share [43] |
| Forward Revenue Multiple | ~16.3x (Analyst Estimate) [44] |
| Forward EBITDA Multiple | ~46.1x (Analyst Estimate) [44] |
The most critical financial driver for valuation is the 5-year sales growth, which is predicated on the successful commissioning of the first 2.0 GW phase by 2027 and a subsequent ramp to 11 GW.[6, 20, 21] If Fermi achieves $125/kW/month in lease rates, 2.0 GW of capacity would yield approximately $3 billion in annual recurring revenue (ARR) with high EBITDA margins.
The investment thesis for Fermi Inc. is highly asymmetric, with risks that are as massive as the potential rewards.
Early Warning Sign: Failure to announce a permanent CEO and a signed, binding lease agreement by the end of Q2 2026.
Thesis Killer: A permanent denial of nuclear licensing or a loss of the TTUS ground lease.[6, 15]
This scenario analysis projects Fermi Inc.’s trajectory through 2031, based on the successful (or unsuccessful) execution of the 11 GW Project Matador campus.
In this scenario, the new leadership successfully stabilizes the company. The first 2.0 GW of gas-fired capacity comes online in late 2027 as targeted. The company secures a mix of neocloud and hyperscale tenants.
* Operational Capacity by 2031: 4.5 GW (Mix of Natural Gas and Grid).
* Lease Revenue Assumption: $110 / kW / Month.
* Annual Revenue (2031): ~$5.94 Billion.
* EBITDA Margin: 60% ($3.56 Billion).
* Exit Multiple: 14x EV/EBITDA.
* Implied Share Price: ~$55.00 (Accounting for dilution to ~900M shares).
* Annualized Return: ~60% from current levels.
Fermi becomes the global standard for private AI infrastructure. Sovereign wealth funds provide massive capital, allowing for an accelerated nuclear and solar build-out. Hyperscalers pay a premium for "sovereign power."
* Operational Capacity by 2031: 8.5 GW (Includes early SMR units).
* Lease Revenue Assumption: $135 / kW / Month (Premium for nuclear/green power).
* Annual Revenue (2031): ~$13.77 Billion.
* EBITDA Margin: 70% ($9.64 Billion).
* Exit Multiple: 18x EV/EBITDA.
* Implied Share Price: ~$165.00 (Reflecting "Super-Utility" status).
* Annualized Return: ~100%+ from current levels.
Management fails to secure a permanent CEO. Litigation and the Fuzzy Panda report scare off major hyperscalers. The company is forced to sell parts of the project at distressed prices to repay equipment loans.
* Operational Capacity by 2031: 0.5 GW (Grid-supply only).
* Lease Revenue Assumption: $85 / kW / Month (Low-end pricing to attract tenants).
* Annual Revenue (2031): ~$510 Million.
* EBITDA Margin: 20% (High overhead and debt service).
* Exit Multiple: 8x EV/EBITDA.
* Implied Share Price: ~$2.50.
* Annualized Return: -14%.
| Scenario | Year 5 Revenue (Est) | Margin / EBITDA | Valuation Multiple | Current Price | Implied Price (2031) | 5-Year Total Return | Annualized Return | Probability |
|---|---|---|---|---|---|---|---|---|
| High | $13.77 B | 70% | 18x EV/EBITDA | $5.20 | $165.00 | +3,073% | 100% | 15% |
| Base | $5.94 B | 60% | 14x EV/EBITDA | $5.20 | $55.00 | +957% | 60% | 50% |
| Low | $0.51 B | 20% | 8x EV/EBITDA | $5.20 | $2.50 | -52% | -14% | 35% |
Probability Weighted Price Target: $53.12
BINARY RISK PROFILE
| Metric | Score (1–10) | Narrative |
|---|---|---|
| Management Alignment | 3 | Former CEO Toby Neugebauer owned 24% of shares, which is high, but the recent abrupt departures and insider selling by Griffin Perry (11M shares) are significant red flags regarding long-term alignment.[17, 31, 51] |
| Revenue Quality | 1 | Currently non-existent. Future revenue (20-year NNN leases) will be high quality, but for a pre-revenue firm, this score remains at the floor.[7, 28, 34] |
| Market Position | 7 | Fermi holds a unique "first-mover" advantage in gigawatt-scale private grids. However, they are currently losing ground in terms of market confidence.[6, 7, 9] |
| Growth Outlook | 9 | The fundamental demand for AI power is unprecedented. The 11-17 GW pipeline represents one of the largest infrastructure targets in history.[6, 13, 20] |
| Financial Health | 5 | Strong cash balance ($408.5M), but massive CAPEX needs ($80B+) and dependence on high-interest equipment loans create a fragile balance sheet.[21, 36, 52] |
| Business Viability | 4 | The "physics" work, but the "execution" is unproven. Reliance on nuclear licensing and one-of-a-kind campus scale creates massive single-point failure risks.[6, 11] |
| Capital Allocation | 6 | Deploying $570M into equipment and permits during the startup phase shows aggressive, milestone-focused allocation, but the "sell-to-cover" activity by current officers is poor signaling.[21, 36, 53] |
| Analyst Sentiment | 6 | Consensus remains a "Buy," but targets are wildly divergent ($8 to $35), indicating a lack of conviction in the models.[37, 54, 55] |
| Profitability | 2 | Currently hemorrhaging cash on a GAAP basis (-$486M loss). Positive cash flow is likely 3–4 years away.[21, 28] |
| Track Record | 2 | As a 2025-founded entity, there is no historical data. The early track record is marred by the loss of an anchor tenant and a management purge.[2, 31, 46] |
OVERALL BLENDED SCORE: 4.5 / 10
SPECULATIVE INFRASTRUCTURE BET
The investment thesis for Fermi Inc. (FRMI) is a structural bet on the physical sovereignty of energy. In a world where the public utility grid is failing to keep pace with the power requirements of artificial intelligence, Fermi’s "behind-the-meter" HyperGrid is a logical, albeit high-risk, solution. The company’s massive 11 GW to 17 GW pipeline and its unique permitting lead in the Texas Panhandle provide it with a significant head start over traditional, grid-reliant data center providers.[3, 6, 7, 13]
However, the "Fermi 2.0" transition marks a critical crossroads. The loss of the founding leadership, combined with the anchor tenant exit and subsequent litigation, has introduced a level of execution risk that cannot be ignored.[14, 31, 46] The next twelve months are critical: the company must prove it can convert "non-binding interest" into long-term, binding contracts and maintain its construction schedule despite these internal shifts.[6, 7, 15] For an investor, Fermi represents a binary opportunity: if it successfully builds the world’s largest private AI grid, the current $5.20 price will appear as a massive valuation gap; if execution continues to slip, the company faces potential insolvency under the weight of its $80 billion CAPEX ambition.[20, 56]
HIGH-STAKES EXECUTION PLAY
Fermi (FRMI) is currently in a state of technical collapse. The stock is trading well below its 50-day moving average of $7.72 and its 200-day moving average of $13.18.[37] The "gap down" to $5.27 on April 20, 2026, following the CFO’s resignation and the Fuzzy Panda short report, has created a significant overhead resistance level.[30, 37, 50] In the short term, the outlook is bearish as the market digests the leadership transition and the ongoing securities litigation.[31, 41] A period of consolidation and a clear permanent CEO announcement are required before a technical floor can be established.
BEARISH MOMENTUM DOMINATES
View Fermi Inc. (FRMI) stock page
Loading the interactive version of this report…