Applied Digital is a high-speed builder of specialized data centers that provide the massive power and liquid cooling required to run artificial intelligence. The business has scaled from just $6.5 million in revenue three years ago to $220 million in its most recently completed fiscal year. It stands today as a primary infrastructure partner for the AI boom, recently securing a $2.15 billion debt raise to fund the construction of massive "AI Factory" campuses in North Dakota.
The investment thesis on Applied Digital is that it owns the rarest resource in the AI economy: shovel-ready data center capacity with hundreds of megawatts of secured power. While tech giants are racing to build AI models, they are hitting a physical wall of power shortages and cooling constraints that only specialized builders can solve. If Applied Digital can execute its massive construction pipeline, it converts speculative infrastructure into a steady, multi-decade annuity from the world's strongest companies.
We think Applied Digital is a high-quality business that has simply become too expensive for new investors at the current price. The company has proven it can land massive, investment-grade tenants, but the stock price has outrun even the most aggressive growth assumptions. We would wait for a significantly better entry point before owning what is otherwise a vital piece of the AI landscape.
Applied Digital stock has soared over the last few years as the company transformed into a major builder of data centers for artificial intelligence. The business grew from almost nothing to a massive operation by securing rare space and power for tech giants. Investors are betting big as the company signs multi-billion dollar building contracts.
What does it do?
Applied Digital is a hypergrowth infrastructure business that earns money by designing, building, and leasing high-performance data centers specialized for artificial intelligence and high-density computing. The company operates as a landlord for the AI era, providing the physical space, massive electrical power, and advanced direct-to-chip liquid cooling that standard data centers lack. Customers sign long-term leases, often spanning 15 to 30 years, and pay a combination of base rent and fees for specialized "fit-out" services. This model turns high upfront construction costs into predictable, long-term cash flows backed by some of the strongest corporate credits in the world.
Where does revenue come from?
Revenue is divided between high-margin AI hosting and its legacy crypto-mining data center business. The HPC Hosting segment, which builds AI Factories, provided $71.0 million in the most recent quarter and is the primary growth engine. The Data Center Hosting segment operates 286 megawatts of capacity for Bitcoin miners, bringing in $37.5 million per quarter. Most of this revenue is generated within the United States, specifically at large-scale campuses in North Dakota.
Who are its customers?
Applied Digital serves a mix of large-scale hyperscalers, specialized AI cloud providers, and Bitcoin mining firms. The company recently signed a 15-year lease for a 200-megawatt campus with a U.S. based investment-grade hyperscaler, which is a massive validation of its infrastructure quality. It also works closely with CoreWeave, an AI cloud leader, providing 250 megawatts across multiple facilities in Ellendale, North Dakota. In its legacy segment, it hosts 286 megawatts of capacity for Bitcoin mining customers across its Jamestown and Ellendale sites. This concentrated group of high-capacity users allows the company to scale revenue rapidly without a large sales force.
What gives it staying power?
The company’s staying power comes from the extreme scarcity of industrial-grade power and the high switching costs of moving thousands of specialized chips. Once a hyperscaler or AI firm installs its billion-dollar server clusters in an Applied Digital facility, moving them is physically difficult and strategically risky.
Where is it headed?
The company is pivoting to become a pure-play AI infrastructure firm by spinning off its cloud services and focusing entirely on building massive 300 to 400-megawatt campuses. Management is betting that the demand for power will far exceed supply for the next decade. If they successfully bring their 1-gigawatt pipeline online, they could reach over $1 billion in net operating income within five years.
Revenue growth is accelerating sharply as massive new data center projects move from construction into active leasing. Revenue reached $126.6 million in the most recent quarter, representing a 139% jump over the prior year. This trend is driven by the first 100-megawatt facility at Polaris Forge reaching full capacity, proving the company can turn its massive physical investments into realized sales.
Cash generation is currently deeply negative as the company spends billions of dollars to build out its 1-gigawatt power pipeline. Free cash flow was negative $800 million in 2025 because the business model requires massive upfront CapEx for land, power lines, and cooling systems years before any rent is collected. This creates a high-risk gap that must be filled by constant debt and equity raises to keep construction moving.
The balance sheet is heavily leveraged with over $2 billion in new debt recently added to fund future growth. While the debt-to-equity ratio of 1.79x is high, the company has successfully secured low-cost financing by using its long-term hyperscaler leases as collateral. This strategy allows them to borrow against guaranteed future rent checks, but it leaves very little room for error if construction projects face delays or cost overruns.
Applied Digital is a hypergrowth infrastructure business that is trading short-term losses for massive long-term contracted revenue.
The first 100-megawatt liquid-cooled data center is now fully operational and generating high-margin rental income. This proves that the company's "AI Factory" design works at scale and can satisfy the technical demands of the most sophisticated AI cloud providers.
The timing of the 200-megawatt Polaris Forge 2 completion is the single biggest risk to the stock. Any delay in bringing this site online would delay the start of its 15-year lease and force the company to carry its $2.15 billion debt load for longer without offsetting cash flow.
The specialized AI data center market is expanding rapidly as AI models require roughly ten times more power per rack than standard cloud apps. The market for high-density hosting is roughly $15 billion today and is on track to exceed $50 billion by 2028 as tech giants outsource the hardest parts of infrastructure. Pricing power is structural because power is now scarcer than the chips themselves, allowing builders to lock in high rental rates. Applied Digital is an emerging leader in this niche, using its early lead in liquid cooling to win contracts that generic data center providers cannot yet handle.
The market is currently in a state of rational competition because demand for power is so much higher than the available supply. Barriers to entry are high, not because of software, but because of the years required to secure utility approvals for 100-megawatt blocks of power. Winning in this industry is a race for power permits rather than a battle over price.
Major competitors like Equinix and Digital Realty have more capital, but they are often slower to retool old facilities for the liquid cooling that AI requires. The most dangerous threat is that hyperscalers like Microsoft or Google choose to build their own power infrastructure directly, cutting out third-party developers entirely. If the largest customers insource the build, Applied Digital's growth runway could shrink overnight.
Applied Digital is currently gaining share as one of the few providers with a live, 100-megawatt liquid-cooled site. The company is successfully moving from crypto-hosting challenger to a Tier 1 AI infrastructure partner.
The primary source of protection is high switching costs created by the physical difficulty of relocating high-performance computing clusters. Once a tenant installs hundreds of millions of dollars in Nvidia GPUs at an Applied Digital site, the cost and downtime of moving to a rival facility are prohibitive. The 15 to 30-year lease terms signed by hyperscalers prove that customers view this as a permanent relationship.
While gross margins are healthy at 27%, the current negative ROIC reflects the massive construction phase. The numbers show a business that is building a moat through physical scale and specialized engineering, rather than one that already owns a dominant market share. The narrow moat is tied to the first-mover advantage in high-density cooling and the multi-year lead in securing power permits.
The moat is currently strengthening as the company stabilizes its first major sites and secures investment-grade tenants. The most important signal of durability is the 15-year lease signed with a hyperscaler, which effectively "de-risks" the next 200 megawatts of capacity.
Delivered the first 100MW liquid-cooled AI facility on schedule.
Raised $2.15B in debt at 6.75% to fund growth.
CEO Wesley Cummins holds a significant stake as co-founder.
Capital Allocation Track Record
Management has demonstrated exceptional vision and timing by pivoting the business from Bitcoin mining to AI infrastructure before the current boom began. Wesley Cummins has shown the ability to raise billions of dollars in capital during a period of high interest rates, which is the oxygen this business model requires to survive. They have successfully moved from a speculative crypto play to a trusted partner for hyperscalers, a transition that many rivals failed to make.
The primary risk is the high degree of key-person dependence on the CEO, whose personal relationships in the power and tech industries drive the company's deal flow. While the appointment of co-founder Jason Zhang as President strengthens the bench, the strategy is still very founder-led and aggressive. There is also a governance concern regarding the complexity of the ChronoScale spin-off and the high debt levels, which leave the company vulnerable to any execution slips.
We expect revenue to grow from $0.4B in FY2026 to $4.2B in FY2031 (~59% CAGR), with EPS growing from $-0.71 to $2.10. Revenue scales as new high-performance data centers come online to meet the massive demand for AI cloud computing. Operating margins improve as the massive upfront costs of building data centers are spread across a growing base of high-paying AI tenants. EPS grows faster than revenue because the business reaches Operating margin expected to reach ~18% by FY2031.
Hyperscalers lease the full 1GW pipeline at high margins. If big tech companies continue to prioritize speed to market over building their own sites, Applied Digital can lock in billions in high-margin, 20-year revenue.
Cost of debt falls as data center assets stabilize. As facilities move from construction to operational, the company can refinance its expensive debt at utility-like rates, boosting net income.
Liquid cooling technology becomes the industry standard for all AI. As chips get hotter, Applied Digital's early expertise in high-density cooling gives it a structural advantage over legacy data center providers.
Construction delays or cost overruns at Polaris Forge 2. If the massive North Dakota sites face delays, the company will struggle to service its $2.15 billion debt load without operational cash flow.
Hyperscalers insource data center construction to save costs. If the largest tech firms find ways to build their own high-density power sites faster, Applied Digital loses its primary customer base.
AI demand stalls before the 1GW pipeline is fully leased. A cooling in the AI investment cycle could leave Applied Digital with massive, half-finished projects and high interest payments.
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 an EV/EBITDA approach projected to FY2027 (the next fiscal year). It fits this business specifically because Applied Digital is a capital-intensive infrastructure provider where GAAP earnings are heavily distorted by depreciation and amortization (D&A) from data center construction; EV/EBITDA is the industry standard for valuing "toll booth" assets like data centers.
The headline math is (FY2027 EBITDA of $325M × 21x) minus $217M in net debt, divided by 285.6M shares, equaling $23. A 21x multiple sits slightly above the data center peer range of 17x to 20x (Equinix at 20x, Digital Realty at 18x) — we apply a premium because of Applied Digital’s pure-play AI workload focus and higher growth trajectory. Our input basis for the $23 fair value is fully consistent with the deterministic engine’s $23 valuation, which accounts for the long-term ramp to FY2031 EPS of $2.10.
A 5-year DCF cross-check produces a fair value of $26 — within 13% of our EV/EBITDA answer of $23, confirming the result. Using the deterministic engine’s FY2031 EPS of $2.10 and a terminal multiple of 20x, the terminal value is $42; discounting this back at a 10% cost of capital yields a present value that aligns with our infrastructure-based model. Both methods suggest that while the long-term cash flow potential is real, the current $46.59 stock price is over-extrapolating that future today.
We're assuming Applied Digital generates an annualized Adjusted EBITDA of $325 million by the end of FY2027. While the company annualized $176 million in its most recent quarter, the ramp-up of the first 100MW at Polaris Forge and the deployment of fit-out services for AI tenants support a doubling of EBITDA as the business moves from construction to hosting.
We're assuming a consolidated net debt position of $217 million after accounting for post-quarter financing proceeds. While the February 2026 balance sheet showed $600 million in net debt, the $382.5 million in financing that closed after the quarter provides a necessary buffer for the high capex requirements of the upcoming fiscal year.
We're assuming the 285.6 million share count remains stable despite the company’s history of using equity for growth. The recent $2.35 billion private notes offering suggests management is pivoting toward debt financing to protect shareholders, though any return to the equity markets would immediately lower our per-share fair value estimate.
The biggest risk is a "funding gap" where the company exhausts its $2.1 billion cash pile before the Polaris Forge campus reaches full high-margin operations. This would force the company into predatory lending or massive share dilution, potentially knocking $10 to $15 off the per-share fair value as the equity is compressed. Watch the "Total Assets" to "Total Liabilities" ratio for any dip below 1.4x.
Bear case ($12): Construction delays at Polaris Forge exceed 6 months, causing a liquidity crunch as debt service costs eat into remaining cash; or Power pass-through arrangements fail to cover rising electricity costs, forcing Adjusted EBITDA margins below 25%.
Bull case ($42): Applied Digital secures an investment-grade credit rating, lowering its interest burden and allowing for faster site expansion; or Contracted backlog converts to operating revenue 20% faster than projected, with NOI exceeding $400 million by FY2028.
Clearthesis wrote this report from 27 sources, including SEC filings, 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 June 23, 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.
The market is leaning bullish because Applied Digital controls the scarce, shovel-ready power infrastructure that hyperscalers need to deploy AI today. By securing massive power grid agreements in North Dakota and landing a multi-billion dollar lease with a major tech giant, the company has effectively transformed from a fringe operator into a critical utility provider for AI hardware.
Skeptics think that Applied Digital is taking on far too much debt to fund its rapid construction expansion. The company is loading its balance sheet with over a billion dollars in secured notes to build facilities before confirming long-term, profitable occupancy for every building, leaving little room for error if the AI infrastructure demand slows.