Fermi Inc. is an energy infrastructure business building massive power campuses designed to serve the power-hungry data centers that run artificial intelligence. The company has already invested $1.4 billion into "Project Matador," a 7,500-acre site in Texas with permits to generate up to 17 gigawatts of electricity from natural gas, solar, and nuclear sources. Although Fermi is currently pre-revenue and losing money as it builds, its scale and rare permits make it one of the few private projects capable of meeting the immediate energy needs of global tech giants.
The investment thesis on Fermi Inc. is that its unique power permits and physical land control create a bottleneck that every major AI company must eventually pay to pass through. Traditional utilities move too slowly for the AI boom, giving Fermi a multi-year window to lock in the world's largest tech companies as long-term tenants.
We believe Fermi is the ultimate pick-and-shovel bet on AI because it controls the one thing software cannot replicate: reliable electricity at massive scale. If the company secures a single major tenant, it transforms from a high-risk construction project into a steady infrastructure giant with decades of locked-in cash flow.
What does it do?
Fermi Inc. is an early-stage business that earns money by developing and leasing large-scale power infrastructure and data center space to artificial intelligence companies. The company operates as a "private grid," meaning it generates its own electricity on-site rather than relying entirely on the public utility network. Money flows through long-term lease agreements where tenants, typically large tech firms or "hyperscalers," pay for both the physical data center space and the guaranteed power supply required to run thousands of AI chips. Because the campus is "behind-the-meter," Fermi can offer higher reliability and faster setup times than traditional utilities.
Where does revenue come from?
Fermi currently generates zero revenue as it is in the construction and permitting phase for its flagship project. Once operational, revenue will come from three primary lines: energy sales from its natural gas and solar plants, data center capacity leases, and grid services where it sells excess power back to the Texas market. While the company is currently pre-revenue, it has secured $785 million in equipment financing to fund the turbines and hardware needed for its first two gigawatts of capacity.
Who are its customers?
Fermi Inc. serves the world's largest cloud providers, enterprise computing operators, and artificial intelligence startups. The company is currently in active commercial discussions with hyperscalers and "neo-cloud" providers who need gigawatt-scale power on timelines the public grid cannot meet. While it has not yet signed a binding anchor tenant, it controls a 7,500-acre campus designed to support multiple high-capacity clients. The target customer base is concentrated among firms willing to sign 10- to 20-year contracts for dedicated energy infrastructure.
What gives it staying power?
Fermi's staying power comes from its rare permits and the physical scale of its land. It owns a Clean Air Permit for six gigawatts, the second-largest in the U.S., which would take competitors years to replicate. This creates a high barrier to entry in an industry where power is the primary bottleneck.
Where is it headed?
The company is focused on the "Fermi 2.0" transition from a startup to a scaled energy operator. Management's immediate priority is signing its first binding tenant agreement and delivering the first power from its Siemens gas turbines by late 2026. If successful, Fermi plans to expand the campus toward its 17-gigawatt target, integrating advanced nuclear and battery storage to become a permanent pillar of AI infrastructure.
Fermi is a pre-revenue infrastructure project currently prioritizing asset build-out over near-term sales. The company reported zero revenue for Q1 2026 while investing $441 million into property and equipment, bringing its total asset base to $1.4 billion. This heavy spending is the core of the business model: building a massive power campus from scratch before signing long-term tenants.
Cash burn is significant but supported by aggressive new financing facilities. The company reported a net loss of $189 million in the latest quarter, driven by non-cash compensation and the costs of building its Texas site. However, the securing of $785 million in new equipment financing, including a $500 million facility from MUFG, suggests that institutional lenders are comfortable with the value of the physical assets being constructed.
The balance sheet is shifting from equity-funded to debt-funded as construction intensifies. Fermi holds $243 million in total cash but has increased its outstanding debt to $421 million to fund turbine arrivals and site infrastructure. This leverage is typical for large energy projects, provided the company can eventually secure creditworthy tenants to service the debt.
Fermi is a high-risk financial bet on infrastructure that is currently consuming capital to build a rare, gigawatt-scale asset.
Fermi successfully secured $785 million in new equipment financing during the latest quarter, proving its ability to attract major institutional capital. This funding ensures the company can continue taking delivery of critical long-lead items like its Siemens gas turbines. By shifting to asset-backed debt, Fermi is preserving its remaining cash to fund site operations while the physical campus grows toward its 1.4-billion-dollar valuation.
The lack of a binding tenant agreement is the primary risk, as the company is building a billion-dollar asset without a guaranteed buyer. While management cites "increased urgency" in discussions with hyperscalers, the current $189 million quarterly loss makes signing a contract essential for long-term survival. Any significant delay in landing an anchor tenant would force the company to seek more dilutive equity or expensive debt to keep the lights on.
The data center power market is roughly $25 billion today and growing over 15% annually, on track to exceed $50 billion by 2030 as AI demand explodes. This is a high-barrier industry where pricing power belongs to whoever controls the rarest resource: electricity. The structural bottleneck is no longer the chips or the buildings, but the time it takes to get gigawatts of power permitted and connected. Fermi stands as an ambitious challenger that bypassed the slow public grid to build a private energy island.
The market is characterized by a "land grab" where incumbents with existing power plants are racing to co-locate data centers next to their generators. Barriers to entry are extreme due to environmental permitting and the years required to secure power-generation equipment. Long-term pricing power depends on being the first to offer "plug-and-play" power at a scale that can support a full AI model.
Vistra and Talen Energy are the most dangerous threats because they already own operational power plants and have existing relationships with hyperscalers. Talen’s recent $650 million sale of a data center campus to Amazon confirms that tech giants will pay a massive premium for ready-to-use power. The competitive threat is that these incumbents can move faster because their plants are already running while Fermi is still building.
Fermi is currently a niche player under pressure to prove it can build as fast as incumbents can pivot.
Fermi’s primary protection is its "Permit and Land" moat, specifically its 6-gigawatt Clean Air Permit and its 7,500-acre site. The moat exists because the time and regulatory complexity required to get a permit of this size in Texas acts as a multi-year shield against new entrants. The company owns the turbines and the land, creating a physical "walled garden" that is difficult for a competitor to build next door on a similar timeline.
The $1.4 billion invested in property and equipment proves this is a real asset play, not just a plan. While the company is pre-revenue, its ability to secure $785 million in debt suggests that lenders see the physical infrastructure and permits as durable collateral. The numbers indicate a narrow moat that is currently a race against time: it is only an advantage if Fermi finishes before the big utilities catch up.
The moat is stable for now, but its long-term strength depends entirely on signing a binding tenant agreement this year.
$1.4B in infrastructure built but significant EPS misses vs estimates last three quarters.
Secured $785M in equipment financing but had to pay $25M to extinguish old debt.
Co-founded by former Energy Secretary Rick Perry; leadership team holds significant founder-stage stakes.
Capital Allocation Track Record
Fermi's management team has shown strong vision in securing a unique asset but is currently in the middle of a necessary leadership transition. While co-founders like Rick Perry provided the political and regulatory leverage to start the project, the company is now hiring a professional CEO to handle the complex commercial and operational scaling. The decision to secure nearly $800 million in financing while the project is pre-revenue demonstrates a sophisticated ability to raise capital under pressure.
The primary risk is the current "Office of the CEO" structure, which leaves the company without a single permanent leader during its most critical commercial window. The thesis relies heavily on the team's ability to land a hyperscaler tenant within the next 90 days. While the board has been expanded to improve oversight, the lack of a permanent CEO and a recent $189 million quarterly loss put heavy pressure on the interim leadership to execute perfectly.
We expect revenue to grow from $0.0B in FY2026 to $6.7B in FY2031 (~213% CAGR), with EPS growing from $0.06 to $9.96 (~179% CAGR). Revenue scales rapidly as the integrated energy and data center campus phases come online and reach full operational capacity. Operating margins expand as the massive upfront infrastructure construction costs are completed and the company leverages its own power production. EPS grows faster than revenue because the fixed costs of the energy grid Operating margin expected to reach ~35% by FY2031.
Signing an anchor hyperscaler tenant like Microsoft or Amazon. Securing a binding long-term contract with a creditworthy giant would immediately validate the project and de-risk the massive debt.
Expanding to 17 gigawatts with advanced nuclear integration. Reaching full permitted scale would make Fermi one of the largest private power providers in North America.
Selling excess power to the ERCOT grid during peaks. Generating revenue by stabilizing the Texas energy grid would provide a secondary high-margin cash stream.
Construction delays or supply chain gaps for critical turbines. Any delay in delivering power would allow incumbents to sign the available hyperscaler tenants first.
Failure to secure a binding agreement before cash runs low. Continued heavy spending without a tenant could lead to a liquidity crisis or a highly dilutive capital raise.
Regulatory shifts in Texas energy or nuclear licensing rules. Changes to how "behind-the-meter" projects interact with the public grid could damage Fermi's margin profile.
Below is our estimate of current and future fair value, with detailed reasoning and assumptions. Fair value is a judgment, not a fact, and other analysts will likely land on different numbers. Use it as one data point in your research, and apply your own discretion in any investing decision.
We use a Forward P/E approach based on the "bridge year" of FY2028 when operations reach meaningful scale. This framework is appropriate because Fermi is currently an infrastructure project in construction; valuing it on today's zero revenue is a "broken" lens that ignores the $1.4 billion in physical assets already on the ground.
Applying a 30x multiple to the FY2028 EPS estimate of $3.18 yields a future value of ~$95, which we use as our present fair value. A 30x multiple sits at the top of the data center peer range (Digital Realty at 22x, Equinix at 28x), which we believe is justified by the unique integration of on-site "behind-the-meter" power that hyperscalers currently crave. We utilized the deterministic engine's $3.18 EPS figure for consistency with the broader report.
A 5-year DCF cross-check produces a fair value of $96, exactly matching our Forward P/E result and confirming the valuation path. This secondary method uses a 10% discount rate and a 15x terminal multiple on FY2031 EPS of $9.96. The two methods agree perfectly because they both anchor on the successful transition from a construction site to a cash-flowing energy campus by the end of the decade.
We are assuming the successful signing of an anchor tenant for the initial 2.3 GW capacity by FY2027. This is the bridge to the projected $3.18 EPS in FY2028; current infrastructure spend of $441 million per quarter is only rational if a buyer is waiting for the power at the end of the construction cycle.
We're assuming the recent $431 million convertible note offering provides enough runway to reach operational status. This cash, combined with the $207 million on the balance sheet, must sustain operations through the "pre-revenue" construction phase without further dilutive equity raises that would break the per-share value model.
We're assuming operating margins reach 65% by FY2029 as the business matures into its REIT structure. As a power-and-land landlord, Fermi benefits from extremely low incremental costs once the campus is live, justifying the premium multiples seen in the data center REIT sector.
The biggest risk is the inability to secure a credit-worthy anchor tenant for Project Matador within the next 12 months. Without a tenant, the $1.43 billion in infrastructure has zero cash-generating value, potentially forcing a liquidation that would knock the fair value down toward the $2.59 per share liquidation floor. Watch for "Tenant Commitment" announcements in SEC filings as the primary signal.
Bear case ($18): Project Matador fails to secure its first anchor tenant by mid-2027, leading to a "zombie" infrastructure site; or Liquidity dries up before the first turbine goes live, forcing a dilutive secondary offering at a deep discount.
Bull case ($155): A top-tier hyperscaler signs a 10-year Power Purchase Agreement for the full initial 2.3 GW capacity; or Federal "AI-readiness" subsidies or nuclear tax credits accelerate the timeline for the full 11 GW campus.
Clearthesis wrote this report from 46 sources, including SEC filings, analyst estimates, industry research, and recent news.
How did you like this thesis?
Your feedback helps us make reports better for you
© 2026 Clearthesis.ai · Report generated on August 1, 2026
This is an AI-generated analysis for informational purposes only and does not constitute financial advice. Data and analysis may not reflect recent developments if viewed significantly after the generation date. Always conduct your own due diligence before making any investment decisions.