APPLIED DIGITAL CORPORATION (APLD): what the price assumes

In the published model solve dated 2026-Q2, anchored at $29.81, APPLIED DIGITAL CORPORATION (APLD) is priced for today's economics sustained for ~40.0 years+. boothcheck publishes no house fair value, target price, or buy/sell rating; individual model outputs and user-controlled scenarios are analytical inputs, not Boothcheck targets. Narrative composed 2026-07-25.

Generated: 2026-08-10 · Exported: 2026-08-11 · Source: https://boothcheck.com/report/APLD

Headline

What The Price Assumes (Inversion)

The assumption today's price embeds, recovered by inverting the valuation.

Beyond 25%/yr sustained for 40 years; not resolvable as a revenue bet. The inversion reports a bound, not a solved point.

Solve inputs: computed at a 17.9% cost of capital; growth searched up to the 25% self-funding ceiling.

Reconcile: at the x-ray's 9.3% required return this reads ~22.4 years; the models below use their own rates.

How unusual the bet is: n/a

Valuation X-Ray

Asset, earnings-power and peer-multiple models all land far below the price; ONLY the growth-DCF reaches it. The bet is durable compounding the static frames structurally cannot price (a moat/durability premium).

How the valuation models price the stock relative to the market price. Price/FV above 1.0 means the market pays more than that lens defends (expensive); at or below 1.0 the lens can defend the price.

Families that justify the price: Growth Families that call it expensive: Asset

The models below discount at their own flat-beta convention rates (cost of equity 9.3%, WACC 5.9%); the inversion above states its own rate.

Per-Model Detail (n=3)

Economic-Unit Decomposition (Sum Of The Parts)

One or more material disclosed units has unresolved economics. Unknown/general is not evidence of homogeneity: segment SOTP is primary but incomplete, consolidated cash flow may remain only a secondary cross-check when every unit shares an enterprise basis, and one sector multiple or target margin is withheld.

No total common-equity value is stated. One or more material units lack a supported unit value. The displayed values are an indicative subtotal; consolidated debt and cash cannot be applied to a fraction of the company.

Solvency

Operating profit is negative or near zero and the company has no demonstrated through-cycle (mid-cycle) operating margin to normalize against, so years-to-repay cannot be computed honestly.

Bullet Takeaways

• Applied Digital is a construction and leasing business wearing a technology label, raising capital to build powered data-center shells in North Dakota and renting them under long-term contracts, a shift the FY2025 10-K describes as the pursuit of "a core specialization in AI focused data centers".
• Tenant concentration is the defining risk rather than merely a listed one: the FY2025 10-K reports two customers accounting for 100% of revenue from continuing operations, one of them at 93% of fiscal 2025 alone.
• Fiscal fourth-quarter and full-year results arrive on July 27, 2026, the first print carrying both a full period of leased AI capacity and the interest expense on the notes issued to build it.

Bull Case

There is no buyback here, no dividend, and no cash held back for optionality. Every dollar this company raises goes into transformers, concrete, and interconnection rights, and the FY2025 10-K says so without varnish: "Construction of our facilities, including the construction of the Polaris Forge 1 campus, are capital-intensive projects, and we anticipate that our current and future strategic growth initiatives will likewise continue to be capital-intensive." Read that as a capital allocation policy and the company stops looking like a technology stock. It is a developer. The question a developer's owner asks is not what the margin was last quarter, but whether the thing being built has a tenant, and whether the money that built it costs less than the rent.

What makes the assets worth something is that the electricity already exists. As of February 28, 2026 the hosting footprint ran a 106 MW facility in Jamestown, North Dakota and a 180 MW facility in Ellendale, per the fiscal 2026 third-quarter filing. Energized, interconnected power at that scale is the scarce input in AI infrastructure at the moment; capital is not. The lease structure then strips out the one cost a landlord cannot control. "Generally, under the terms of our leases, the majority of our rental expenses, including power and security cost, are recovered from our customers." In plain terms, the tenant carries the power price. That turns what would otherwise be a commodity exposure into a spread between rent collected and interest owed.

The tenant matters more than the building. CoreWeave anchors Polaris Forge 1, and CRWV reported trailing revenue of $6.23 billion expanding about 130% year over year at an operating margin near -2.6%, a customer with real scale and real fragility at the same time, whose own 10-K describes deploying "the industry's most advanced hardware and architectures first-including NVIDIA GB200 and GB300 NVL72 systems". The June 2026 memorandum of understanding to assign the Building 3 lease to a CoreWeave subsidiary, conditioned on that subsidiary reaching an investment-grade credit rating, is the part worth watching. If the condition is met, the rent stops depending on a fast-growing company's balance sheet and starts depending on an investment-grade one. That is an upgrade in the quality of the cash flow, not a press event.

How the building gets paid for has changed, and that change is the strongest thing in the bull case. A year ago the funding mechanism was equity: the FY2025 10-K discloses a sales agreement under which up to 200 million dollars of common stock could be issued. In 2026 the company sold secured paper instead, 2.15 billion dollars of 6.750% senior secured notes priced in March and 1.59 billion dollars of 7.000% senior secured notes closed in June, both raised at project subsidiaries against contracted lease cash flows. Credit investors underwriting a decade of rent at those coupons did work the equity market cannot do from outside. They read the leases.

Bear Case

The variable with the most leverage on this equity is not AI demand. It is the price of credit. This company borrows at fixed rates to build assets it then leases out, and the equity is whatever is left between rent received and interest paid. The two 2026 offerings carried coupons of 6.750% and 7.000%, which is what a developer with no operating profit pays even in a receptive market. A developer in this position has to keep coming back to that market. If spreads widen, or if the build cycle simply pauses for two quarters, the next tranche prices worse and the arithmetic that makes a new building worth starting stops working. Very little in today's price suggests that risk is being charged for.

Underneath the financing sits a customer list short enough to fit on a napkin. The FY2025 10-K states "customers accounting for 100% of our revenue from continuing operations. One of such customers accounted for 93% of our revenue during fiscal year 2025." The legacy hosting segment ran on a single crypto-mining tenant with, at the time of that filing, a remaining contractual term of two and a half years. The risk factors do not pretend the next building is spoken for either: "We cannot assure you that once we have developed space or land we will be able to successfully lease it at all, or at rates we consider favorable". A vacant powered shell still owes interest.

Then there is what the price itself demands. At roughly 17 times revenue against a business running a trailing operating margin of -42.6%, the price is asking for a transformation rather than an improvement. Push revenue higher by 25% a year, the ceiling this business could fund out of its own cash flows, hold that pace for 40 years, and the price still does not resolve into any revenue path that reaches it. That is a harder statement than "expensive." The persistence being asked for is itself uncommon: among comparable fast climbers, only about 34% held that pace even five years. And if the requirement reverts toward what the sector actually delivers, there is a long way down before the price meets a standard frame, because it currently sits about 9.7 times above where peer multiples land, 5.1 times above the asset-value methods, and 5.0 times above the forward-growth methods.

The last piece is what has already been taken from existing holders. Share count has compounded at 51.6% a year over the four years to February 28, 2026. Some of that bought megawatts that now produce rent, which is the bull's honest answer. But the mechanism does not distinguish between good projects and bad ones, and it runs constantly: the FY2025 10-K describes a preferred issue raising $150.0 million in gross proceeds, of which 156,000 preferred shares later converted into roughly 21.0 million common shares, with the equity sales agreement sitting behind it. An owner who liked every project still owns a far smaller fraction of them than when they started.

Valuation

Start with the mechanical problem in pricing this company: there is no operating profit to price. The trailing operating margin is -42.6%, so the market is valuing sales instead, and at $27.20 the shares carry about 17 times revenue. That top line is worth understanding before the multiple is. The fiscal 2026 third-quarter filing records amounts reimbursable by customers, its "tenant recoveries", as revenue in the period the underlying expenses are incurred, so recovered power cost sits inside the denominator of that multiple alongside the rent itself. What the multiple embeds on the other side is an eventual operating margin near 7.9%, roughly what this cost structure would convert to once the buildings are full and the construction spending stops. The distance between -42.6% and 7.9% is the whole question.

The more revealing exercise is to ask what path of revenue would justify the price, and to find that none does. Hold revenue expanding at 25% a year, as fast as this business could go without raising fresh money, run it for 40 years, and discount at the 17.7% cost of capital these cash flows carry, and the number still does not arrive. That is not a demanding assumption that may or may not come true. It is a requirement the arithmetic cannot satisfy on the inputs the business currently shows, which says the price is being set by something other than an extrapolation of the existing operations.

Every family of method lands below the price, which is itself unusual. Most stretched names have at least one lens that reaches. Peer multiples are the furthest away, with the price sitting about 9.7 times above where that lens lands; the asset-value methods, which read the balance sheet and the returns earned on it, sit about 5.1 times below, and the forward-growth methods about 5.0 times below.

The friendliest construction available is a forward market-value method. It extends revenue at the pace the company has recently managed, tapers that pace as the base grows, and holds today's revenue multiple flat at the end of the horizon, compressing it in the bear scenario and expanding it in the bull. Even on those terms it lands under the price. When the most generous arithmetic on the board still falls short, the gap is not a modelling artifact.

The cohort supplies scale rather than comfort. CRWV carries trailing revenue of $6.23 billion at an operating margin near -2.6%, and IREN $1.07 billion at -39.6%, both building the same kind of capacity into the same demand. ORCL, the only member of the set earning money at the operating line, runs a 30.6% operating margin on $67.4 billion of revenue, a reminder of what mature infrastructure economics eventually look like. Applied Digital's revenue base is far smaller than any of them, while its market value assumes it will not stay that way.

The balance sheet comes last, and it needs a date attached. At February 28, 2026 the company held more liquid assets than funded borrowings at the corporate level, which reads comfortably until you notice what followed: 2.15 billion dollars of secured notes in March and 1.59 billion dollars more in June, both drawn at project subsidiaries against the leases. Interest coverage stands at -7.2 because there is no operating profit to cover interest with, and the coupons therefore get paid out of financing proceeds and tenant rent rather than out of earnings. Cash is being consumed, not generated. What bounds the downside is not the balance sheet, then, but the leases behind those notes, and a 7.7 billion dollar market value resting on a hosting business the FY2025 10-K describes as having one customer is a narrow foundation by any reading.

Catalysts

Fiscal fourth-quarter and full-year 2026 results are due after the close on July 27, 2026, with the call at 5:00 p.m. Eastern. The fiscal fourth-quarter revenue consensus stands at about $82 million, roughly 116% above the same quarter a year earlier. Revenue and the pace of energization are the lines that matter on that print, because together they determine when leased capacity starts paying against the interest that funded it.

The financing calendar has been the real news of 2026. In March the company priced 2.15 billion dollars of 6.750% senior secured notes due 2031 through a project subsidiary, funding 200 MW at the Polaris Forge 2 campus in Harwood, North Dakota. In June it closed a further 1.59 billion dollars of 7.000% senior secured notes due 2031 to build ELN-04, a 150 MW facility at Polaris Forge 1 in Ellendale. Each raise converts directly into contracted capacity. Each one also fixes an interest bill that arrives on schedule whether or not the building does.

Two operating milestones frame the next twelve months. Polaris Forge 1 reached ready-for-service on its first building, delivering 100 MW to the anchor tenant on time, the first of three contracted buildings on a campus planned at 400 MW. And in June the company signed a memorandum of understanding to assign the Building 3 lease to a CoreWeave subsidiary, contingent on that subsidiary obtaining an investment-grade credit rating. Watch that condition. Whether it is satisfied changes the credit quality standing behind a large share of the rent, and for this equity that is a bigger swing factor than any single quarter's revenue.

Peer Cohorts (Per Segment, With Filing Citations)

Core business (reported)

• CRWV (CoreWeave Inc)
• FY2025 10-K: …while improving efficiency and lowering total cost of ownership. Our track record and ability to 6 Table of Contents deploy the industry's most advanced hardware and architectures first-including NVIDIA GB200 and GB300 NVL72 systems for mission-critical AI-gives our customers a measurable edge in performance,…
• FY2025 10-K: …satisfactory to us when we require it, our ability to continue to support our business growth and to respond to business challenges could be significantly impaired and our business may be adversely affected, requiring us to delay, reduce, or eliminate some or all of our operations. Even if we are able to raise such…
• NBIS (Nebius Group NV)
• (no filing in the citation store)
• IREN (IREN Ltd)
• FY2025 10-K: …Our goal is to maximize utilization of our AI Cloud Services capacity by customers. We aim to enter into contracts for our AI Cloud Services once the relevant GPUs have been ordered or are installed and operational, and if any such computing power becomes available upon expiration of a customer contract or otherwise.…
• FY2025 10-K: …in which we operate (for example, through potential participation in demand response, ancillary services provision and load management in deregulated markets such as Texas). We have secured sites with access to land and power supply, which we believe positions us to take advantage of any growth in power demand for…
• CORZ (Core Scientific Inc)
• FY2025 10-K: …for certain facility-level operating expenses, while the customer is billed separately for power and, in some cases, other variable operating costs. Both lease 8 Table of Contents structures are frequently paired with long-term, take-or-pay contractual commitments, under which the customer is obligated to pay for…
• FY2025 10-K: …intelligence and HPC workloads. We plan to develop and bring online the infrastructure required to meet our existing contractual commitments to our high-density colocation customer, expand our infrastructure portfolio by securing additional land and power at new and existing sites, and sign additional colocation…
• ORCL (Oracle Corp)
• FY2025 10-K: …development tools, among others. These infrastructure technologies are available through a subscription to our OCI offerings or through the purchase of a license and related license support, at the customer's option, to run within the Oracle Cloud as a part of a customer's cloud-based, on-premise or other IT…
• FY2025 10-K: …such as in the Oracle Exadata Cloud@Customer deployment models (described further below). We focus on the operation and integration of Oracle infrastructure technologies to make them easier to deploy, extend, interconnect, manage and maintain for our customers and to improve computing performance relative to our…

Methodology Note

• Priced-in inversion: the valuation is inverted on the current price to recover the operating-income growth, duration, and steady-state margin the price embeds (ROE for financials, FFO growth for REITs).
• Valuation x-ray: the valuation models, grouped into four families (asset, earnings, relative, growth). Each model is expressed as a price/FV ratio (distance from price), not a point fair-value estimate. The spread across families is the disagreement.
• Economic-unit decomposition (SOTP): each disclosed business unit is assigned its native valuation basis before any multiple is applied. Operating units are valued on enterprise value; funded financial units are valued on their own common equity, because their borrowings fund earning assets rather than levering the parent. A company total is stated only once every material unit carries a supported value and the parent capital bridge reconciles.
• Solvency: net cash/debt, net-debt-to-NOPAT, interest coverage, and share-count CAGR from EDGAR financials (net debt / FFO and fixed-charge coverage for REITs; regulatory-capital framing for financials).
• Peer cohorts: per-segment comparables with deep-linkable SEC filing citations.

Fundamentals sourced from SEC EDGAR filings. Current price from Databento. The priced-in inversion and valuation x-ray are computed by the boothcheck engine; narrative composed by AI from the structured data.