Applied Digital Corp.

Executive Summary

BUY, Medium conviction. Rated BUY, Medium conviction for investors able to underwrite construction financing and lease execution. The first 100 MW online produced a sharp Q4 earnings inflection, while the stock sits 58.3% below its 52-week high and below the 200-day average. The key question is not demand, which management says is strengthening, but whether APLD can fund and deliver its contracted campuses without further damaging per-share economics.

Investment Thesis

The market is treating APLD as a capital-hungry developer, and that skepticism is justified. The company carries a balance sheet built for construction, not a mature landlord, and its reported income statement is still deeply negative. But the share price has already absorbed a large amount of that concern. At $21.16, investors are paying for a business that has only begun to show the economics of its contracted AI infrastructure rather than for a fully proven recurring-NOI platform.

The investable change is operational. The first online capacity moved the company from a negative Q3 adjusted EBITDA margin to a positive Q4 outcome, and the underlying rental model produced a 91% NOI margin. That does not make the equity safe. It does establish that the completed assets can be structurally different from the consolidated company that bears start-up cost, depreciation, stock compensation and construction spending. The investment case depends first on additional megawatts entering service, second on the 250 MW of expansion discussions becoming signed leases, and third on financing remaining available at a cost that does not erase the value created by those assets.

The stock is not a conventional value purchase. There is no usable trailing P/E and no disclosed forward EPS estimate against which to claim a low forward multiple. The appropriate underwriting framework is enterprise value per contracted and operating megawatt, recurring NOI, and the equity residual after debt. That framework supports a BUY only for investors who accept that a delay in power delivery or lease commencement can have a disproportionately negative equity impact.

Management's latest EPS beat is a positive inflection, not proof of repeatability after several recent misses. Medium conviction reflects that distinction. The upside is meaningful if AI demand continues to reward scarce powered capacity and APLD delivers as scheduled; the downside remains substantial if its build schedule gets ahead of its funding capacity.

Key Metrics

Financial Strength

APLD has sufficient reported near-term liquidity through a 2.56x current ratio, but its financial profile is defined by leverage-funded construction rather than internally financed growth. The recent financings include $2.15 billion of 6.75% notes, $1.59 billion of 7.0% notes, a $300 million bridge and up to a $550 million revolver. Those facilities provide build capacity, yet they also raise the hurdle for equity value creation. Margin trajectory has improved materially at the project level, but the balance sheet will remain fragile until recurring cash rent covers corporate costs, interest and ongoing development needs.

Competitive Position

APLD's near-term protection comes from contracted power-intensive AI campuses, embedded hyperscaler relationships, and the difficulty of securing power, land, equipment and financing simultaneously. The $36 billion contracted lease portfolio gives it more demand visibility than an unleased speculative build. Durability is not yet proven: larger data-center operators and hyperscalers have superior capital access, and APLD's advantage will depend on completing power-connected sites before customers can source equivalent capacity elsewhere.

Management & Guidance

Management is not offering conventional revenue or EPS guidance. Instead, it is guiding investors to roughly $600 million of next-quarter CapEx, a potential 250 MW of incremental expansion capacity, and a $1 billion NOI run-rate within four years. CEO Wesley Cummins stated, "Finally, a year ago, we set a goal for $1 billion of net operating income within 5 years. We now expect to achieve that run rate goal a year from now or 3 years ahead of schedule." The guidance is supported by a visible contracted portfolio and improving rental rates, but its credibility is moderated by the recent sequence of EPS misses and by construction dependencies outside management's direct control, especially power and supply chain.

Segment Analysis

• HPC and AI data-center capacity (Not disclosed): This is the strategic growth engine, with initial rent from Polaris Forge 1 and tenant fit-out services driving the Q4 step-up.
• Legacy cloud and data-center services (Not disclosed): Product-level revenue and geographic revenue splits were not supplied, preventing a reliable segment-margin or regional concentration analysis.

Capital Allocation

Capital allocation is overwhelmingly directed to construction rather than shareholder distributions. The deep-dive feed records $2.1 billion of CapEx, $0.0 billion of buybacks and $0.0 billion of dividends, appropriate for a capacity buildout but unfavorable for per-share protection. That feed also lists $2.0 billion of FCF, which conflicts with the supplied negative $4.5 billion LTM FCF figure; until the company reconciles cash-flow definitions and converts projects into cash generation, capital discipline cannot be judged favorably.

Management Execution & Track Record

A clean two-year promise-versus-delivery comparison is not possible because the supplied 2025 Q1 call is marked stale and contains no guidance, margin, cash-flow, or project-timeline benchmarks. The current record is mixed: FY2026 revenue grew 33.6%, operating margin improved from negative 31.6% to negative 28.0%, and Q4 adjusted EBITDA rose to $42.4 million, demonstrating meaningful operating progress. However, the company missed EPS expectations in five of the subsequent eight quarters and posted a $111.6 million latest-quarter net loss, largely tied to ChronoScale-separation and PSU-related stock compensation. Management's claim that projects are on time and on budget and that the $1 billion NOI objective is now a year earlier than planned is encouraging, but credible execution must be demonstrated through signed expansions, delivered megawatts and lower cash burn.

Positioning & Flows

Institutional ownership is concentrated, with the top three institutions holding 60.3% and insiders holding 7.6%. Recent insider purchases included $6.7 million in aggregate disclosed buying, which aligns management with a depressed share price but does not offset financing risk. Short interest is elevated at 56.6 million shares, or 21.66% of shares with 2.9 days to cover; positive gamma positioning and a $20.10 zero-gamma level, 5.17% below spot, may limit immediate downside but can also amplify volatility around earnings.

Bull Case

• Q4 adjusted EBITDA margin reached 16.4%, improving 2,042 bps from Q3's negative 4.0% as the first 100 MW began contributing.
• The HPC base-rental model generated a 91% NOI margin, indicating that completed and occupied campuses can earn far better economics than the company's reported consolidated margins suggest.
• Management has five AI factory campuses under construction for three hyperscalers, reducing customer-acquisition uncertainty versus a speculative data-center buildout.
• AI capacity demand is translating into pricing power: management reported rental rates increasing over the past six months while marketing 1.7 GW of new capacity.

Bear Case

• Total debt was $6.5 billion as of 2026-08-31 and debt-to-equity was 3.63x, leaving equity holders exposed if construction costs, lease timing, or financing terms worsen.
• LTM free cash flow was negative $4.5 billion through 2026-08-31, and management expects roughly $600 million of CapEx in the upcoming quarter before further increases at newer campuses.
• FY2026 operating margin remained negative 28.0% and net margin was negative 110.0%, so the Q4 improvement has not yet established sustainable corporate profitability.
• Earnings execution remains uneven: APLD missed Street EPS in five of the past eight reported quarters, including misses of 260.0% and 242.1% in April and July 2026.

Valuation & Price Target

Engine price target $25.59 (+20.9% vs current), High confidence.

• P/S Multiple: $27.48 (45% weight)
• Quality-adjusted: $21.33 (20% weight)

The quantitative engine's HOLD, High conviction call is reasonable because it assigns a $25.59 target, only 20.9% above spot, and its quality score is just 51/100 against a moderate 45/100 risk score. This note moves to BUY because the engine's blended values span $21.33 to $27.48 and appear to underweight the step-change from Q4 EBITDA of $42.4 million, contracted lease value, and the stock's severe reset from its $50.73 high. Conviction remains Medium because leverage and negative free cash flow can overwhelm that upside. Sell-side Strong Buy consensus, based on seven buys and one hold, is low-signal context rather than support for the rating.

Risk Assessment

Risk score 45/100 (Moderate). The central risk is that $6.5 billion of debt and negative $4.5 billion LTM free cash flow force dilutive capital raises or expensive refinancing before contracted projects generate their expected NOI.

The Bottom Line

APLD rates a BUY with Medium conviction for the next 6-12 months. The valuation offers upside if contracted AI capacity becomes operating capacity and the business continues shifting toward high-margin rental NOI. This is not a low-risk AI proxy: debt, free-cash-flow burn and inconsistent earnings delivery make position sizing essential. The rating would move to HOLD if January results fail to show lease conversion or if project funding requires material incremental dilution; it would strengthen if management signs the 250 MW expansions and demonstrates recurring cash generation.