Fermi Inc. is developing a private power campus for AI-centric customers, focusing on grid-independent energy generation and high-performance computing facilities designed for the hyperscale era. The company aims to deliver low-carbon, highly reliable, and on-demand power to compute-intensive businesses, leveraging its Project Matador campus in the Texas Panhandle region. Fermi Inc. plans to support up to approximately 15 million square feet of AI-ready hyperscale compute…
Fermi Inc. is developing a private power campus for AI-centric customers, focusing on grid-independent energy generation and high-performance computing facilities designed for the hyperscale era. The company aims to deliver low-carbon, highly reliable, and on-demand power to compute-intensive businesses, leveraging its Project Matador campus in the Texas Panhandle region. Fermi Inc. plans to support up to approximately 15 million square feet of AI-ready hyperscale compute infrastructure through a multi-decade buildout, integrating on-site natural gas-fired generation, battery energy storage systems, solar generation, and longer-term nuclear baseload supply.
Fermi Inc. intends to generate substantially all of its revenue from long-term tenant lease arrangements at Project Matador, with tenant payments structured as rents from real property for U. S. federal income tax purposes. Revenue components include fixed monthly rent based on reserved power capacity expressed in dollars per kilowatt per month, fixed rent based on invested capital in real property improvements, pass-through reimbursements for energy-related input costs and other shared campus system expenses, and development and administrative fees for certain deployments. The company expects tenant payments to be primarily tied to contracted power capacity and service levels, reflecting the economics of high-density AI workloads.
The company operates through the following segments:
• Powered Shell Leasing: This segment involves the development and leasing of powered shell buildings designed to support tenant mechanical, electrical, and plumbing installation and high-density compute deployment. The company plans to construct and deliver initial powered shell facilities including structural buildout, electrical backbone infrastructure, and cooling systems. Fermi America, HyperRedundant, and HyperGrid are trademarked names associated with this segment's branding and service offerings.
• Ground Leasing: This segment offers long-term ground leases to tenants that elect to construct their own powered shell facilities on Fermi Inc.'s land, with or without integrated power delivery. Under this model, the company expects to earn market-based ground lease rent under long-duration lease structures, with power capacity and infrastructure made available under the lease structured as rent or additional rent. The Ground Leasing segment utilizes the same trademarked identifiers—Fermi America, HyperRedundant, and HyperGrid—for consistency in brand representation across tenant-facing services.
Fermi Inc. is positioned to address growing demand for private power solutions in the AI infrastructure market, where traditional utility power delivery faces constraints from limited grid capacity, interconnection queue delays, and transmission congestion. The company's competitive advantages include large-scale site control through a 99-year ground lease with the Texas Tech University System, proximity to productive natural gas-producing regions, access to major fiber optic transmission corridors for low-latency connectivity, and ownership of groundwater rights from the Ogallala Aquifer. These factors support redundant fuel supply, scalable water resources, and high-security infrastructure due to proximity to the U. S. government's Pantex facility.
Fermi Inc.'s target customers are hyperscale and other large-scale compute users, including cloud service providers, AI infrastructure operators, chipmakers, and enterprises or governmental and quasi-governmental entities requiring long-duration access to high-density compute capacity supported by reliable power. The company expects tenant lease agreements to be long-term in nature and structured to support large, phased deployments over time, with early discussions underway with multiple prospective tenants across various stages of negotiation, including those evaluating term sheet structures and conducting confirmatory site diligence.
Sectors:Real Estate · UtilitiesSector rationaleThe company's primary revenue model is based on long-term tenant lease arrangements, specifically 'Powered Shell Leasing' and 'Ground Leasing,' where payments are structured as rents from real property. While it integrates energy generation (natural gas, solar, nuclear) to support these sites, this activity serves as the utility infrastructure for its real estate offerings, justifying a secondary sector classification in Utilities.Industries:+1 moreData Center REITsReal EstatePrimaryFermi develops and leases 'powered shell buildings' and high-performance computing facilities specifically for hyperscale AI customers. Its revenue model is based on long-term tenant lease arrangements for these data-center-ready facilities.Independent Power ProducersUtilitiesSecondaryThe company provides grid-independent energy generation via on-site natural gas-fired generation and battery storage to its tenants. This power is sold as part of the lease structure based on reserved power capacity (dollars per kilowatt per month) outside of a regulated utility rate base.Renewable Power ProducersUtilitiesSecondaryFermi integrates solar generation into its power campus to provide low-carbon energy to its compute-intensive business customers.Classified using BQ-MICSCIK: 0002071778
Investment Thesis
▲ Bull case
Fermi's aggressive infrastructure deployment, with over $1.4 billion already invested in physical assets at Project Matador, positions it to capitalize on the imminent surge in AI-driven power demand from hyperscalers and enterprise compute operators, particularly as the company's 6-gigawatt air permit and ongoing pursuit of an additional 5 gigawatts create a unique, near-term advantage in securing binding tenant agreements that will unlock phased revenue streams starting in 2027, with management's emphasis on tenant diversity and modular MEP partnerships indicating a sophisticated approach to optimizing load profiles and accelerating time-to-power for customers, directly addressing a critical bottleneck in the AI infrastructure value chain that competitors cannot replicate due to Fermi's integrated site control and regulatory progress.
The company's disciplined capital allocation strategy, which ties all future construction spending to definitive tenant agreements and project financing closure, demonstrates prudent risk management while preserving liquidity for at least 12 months, and the recent securing of $885 million in nonrecourse equipment financing facilities—including the $500 million MUFG turbine warehouse and $165 million Yellowstone facility—provides a de-risked path to commissioning core power generation assets without diluting existing shareholders, thereby enabling Fermi to maintain balance sheet strength while advancing toward revenue-generating operations, a nuance underscored by the CFO's clarification that these financings remain off-balance sheet and are structured to align with milestone-based capital deployment.
Fermi's evolution into Fermi 2.0, marked by strengthened governance through bylaw amendments requiring 70% shareholder approval for board changes and the independent committee's rejection of the former CEO's special meeting solicitation, has effectively mitigated governance overhang and refocused the organization on executing its long-term industrial strategy, as evidenced by the Board's endorsement of Fermi 2.0's milestones—including securing binding tenant agreements, managing working capital, and delivering power—and the reaffirmed support from major shareholders like Caddis Capital, which collectively signal a stable, shareholder-aligned leadership committed to realizing the project's 11-gigawatt permitted capacity and eventual 17-gigawatt build-out potential.
Fermi's aggressive infrastructure deployment, with over $1.4 billion already invested in physical assets at Project Matador, positions it to capitalize on the imminent surge in AI-driven power demand from hyperscalers and enterprise compute operators, particularly as the company's 6-gigawatt air permit and ongoing pursuit of an additional 5 gigawatts create a unique, near-term advantage in securing binding tenant agreements that will unlock phased revenue streams starting in 2027, with management's emphasis on tenant diversity and modular MEP partnerships indicating a sophisticated approach to optimizing load profiles and accelerating time-to-power for customers, directly addressing a critical bottleneck in the AI infrastructure value chain that competitors cannot replicate due to Fermi's integrated site control and regulatory progress.
The company's disciplined capital allocation strategy, which ties all future construction spending to definitive tenant agreements and project financing closure, demonstrates prudent risk management while preserving liquidity for at least 12 months, and the recent securing of $885 million in nonrecourse equipment financing facilities—including the $500 million MUFG turbine warehouse and $165 million Yellowstone facility—provides a de-risked path to commissioning core power generation assets without diluting existing shareholders, thereby enabling Fermi to maintain balance sheet strength while advancing toward revenue-generating operations, a nuance underscored by the CFO's clarification that these financings remain off-balance sheet and are structured to align with milestone-based capital deployment.
Fermi's evolution into Fermi 2.0, marked by strengthened governance through bylaw amendments requiring 70% shareholder approval for board changes and the independent committee's rejection of the former CEO's special meeting solicitation, has effectively mitigated governance overhang and refocused the organization on executing its long-term industrial strategy, as evidenced by the Board's endorsement of Fermi 2.0's milestones—including securing binding tenant agreements, managing working capital, and delivering power—and the reaffirmed support from major shareholders like Caddis Capital, which collectively signal a stable, shareholder-aligned leadership committed to realizing the project's 11-gigawatt permitted capacity and eventual 17-gigawatt build-out potential.
Despite Fermi's progress in securing permits and equipment, the absence of any binding tenant agreement as of the May 2026 earnings call, coupled with management's deliberate avoidance of providing timelines for deal execution due to concerns about altering negotiation dynamics, creates significant uncertainty around when revenue will commence, and the company's acknowledgment that tenant revenues will not be sufficient to fund full operating capital requirements until Matador is built out and operating at scale implies a prolonged period of negative cash flow and continued reliance on external financing, which could strain liquidity if project financing is delayed or secured on unfavorable terms, particularly given the company's reliance on future cash flows from tenant commitments to underpin project-level debt.
The company's heavy dependence on securing investment-grade counterparties willing to wrap their balance sheets around Fermi's power offerings introduces a material risk, as highlighted by the CFO's emphasis on this requirement as non-negotiable and the CEO's admission that tenants want all available power forever, which conflicts with Fermi's need for multiple tenants to achieve load diversity and genset efficiency, suggesting that negotiations may stall if hyperscalers insist on exclusive offtake arrangements or if modular MEP partners fail to accelerate their own deployment timelines, thereby creating a mismatch between Fermi's power readiness and tenants' ability to consume it, a dynamic underscored by the CFO's surprise at needing to conduct reverse due diligence on tenants' MEP readiness.
Fermi's aggressive capital deployment to date, with $570 million invested in construction in progress and plans to exceed $3 billion in total aggregate capital for Phase 0 and Phase 1, raises concerns about potential overbuilding in a market where AI infrastructure demand may not materialize as rapidly or at the scale anticipated, especially given the company's admission that it does not expect to pay dividends until REIT taxable income requires it and its history of significant noncash charges—such as the $174 million charitable contribution of Class B units and $61 million in fair value losses on convertible notes—which obscure true economic performance and could mask underlying profitability challenges even after revenue begins, particularly if tenant pricing remains static as indicated by the CFO's comment that pricing is 'at the same' level as initial expectations, limiting upside potential in a potentially competitive power market.
Despite Fermi's progress in securing permits and equipment, the absence of any binding tenant agreement as of the May 2026 earnings call, coupled with management's deliberate avoidance of providing timelines for deal execution due to concerns about altering negotiation dynamics, creates significant uncertainty around when revenue will commence, and the company's acknowledgment that tenant revenues will not be sufficient to fund full operating capital requirements until Matador is built out and operating at scale implies a prolonged period of negative cash flow and continued reliance on external financing, which could strain liquidity if project financing is delayed or secured on unfavorable terms, particularly given the company's reliance on future cash flows from tenant commitments to underpin project-level debt.
The company's heavy dependence on securing investment-grade counterparties willing to wrap their balance sheets around Fermi's power offerings introduces a material risk, as highlighted by the CFO's emphasis on this requirement as non-negotiable and the CEO's admission that tenants want all available power forever, which conflicts with Fermi's need for multiple tenants to achieve load diversity and genset efficiency, suggesting that negotiations may stall if hyperscalers insist on exclusive offtake arrangements or if modular MEP partners fail to accelerate their own deployment timelines, thereby creating a mismatch between Fermi's power readiness and tenants' ability to consume it, a dynamic underscored by the CFO's surprise at needing to conduct reverse due diligence on tenants' MEP readiness.
Fermi's aggressive capital deployment to date, with $570 million invested in construction in progress and plans to exceed $3 billion in total aggregate capital for Phase 0 and Phase 1, raises concerns about potential overbuilding in a market where AI infrastructure demand may not materialize as rapidly or at the scale anticipated, especially given the company's admission that it does not expect to pay dividends until REIT taxable income requires it and its history of significant noncash charges—such as the $174 million charitable contribution of Class B units and $61 million in fair value losses on convertible notes—which obscure true economic performance and could mask underlying profitability challenges even after revenue begins, particularly if tenant pricing remains static as indicated by the CFO's comment that pricing is 'at the same' level as initial expectations, limiting upside potential in a potentially competitive power market.