02 Executive Summary

Applied Digital Corporation is one of the fastest-growing purpose-built AI infrastructure platforms in North America, pivoting from crypto-focused HPC hosting into a hyperscaler-grade data center developer and operator. Revenue has grown from $55.4M in FY2023 to a TTM figure of $319.3M — a 5.8x expansion in roughly three years — while the company simultaneously raised its gross margin from near-zero to 45.4% TTM, signaling a fundamental business model upgrade.

Investment Thesis: APLD is transitioning from a capital-light crypto hosting business into a hyperscaler-grade AI data center developer with long-term leases providing revenue visibility. The signing of a major hyperscaler lease for the Polaris Forge 3 campus confirms that institutional-grade customers are underwriting APLD's build-out thesis.^[S9] At 46.9x EV/Revenue, valuation is demanding but defensible given 139% YoY revenue growth and an emerging contracted backlog.

Top Catalysts:

• Polaris Forge 3 Hyperscaler Lease: A major unnamed hyperscaler signed a long-term lease on APLD's new Polaris Forge 3 campus, extending the contracted revenue model beyond existing facilities and providing multi-year revenue visibility.^[S9]
• EPS Beat Streak: APLD has beaten consensus EPS estimates in each of the last four consecutive quarters, with the FQ4 FY2026 (Feb 2026) quarter delivering a +$0.09 actual vs. -$0.21 estimate — a 142.9% beat — signaling accelerating operating leverage.
• $300M Bridge Facility: The May 2026 announcement of a $300M senior secured bridge facility provides near-term liquidity runway to fund the aggressive CapEx program without immediate equity dilution.^[S16]

Key Risks:

• Negative FCF Burn: TTM FCF stands at -$1.50B, driven by $775.2M in Q3 FY2026 capex alone; sustained capital access is non-negotiable to the thesis.
• Customer Concentration: The hyperscaler-dependent model means lease cancellation or renegotiation by one anchor tenant could materially impair projected revenues.
• Execution Risk at Scale: APLD has never operated data centers at the scale currently under development; construction delays, cost overruns, or commissioning failures could defer revenue recognition.

Our $62.00 price target is derived from a blended 50% DCF base case (12% WACC, 3% terminal growth) and 50% forward EV/Revenue peer multiple of 20x applied to FY2027E revenue of $712.5M. At $45.14, the stock trades at a 35.4% discount to the consensus mean target of $61.10, with 10 analysts covering and zero sell ratings.

03 Financial Performance & Health

3a. Income Statement Analysis

APLD's revenue trajectory is among the most dramatic in its peer group: from $8.5M in FY2022 to $144.2M in FY2025, with TTM revenue reaching $319.3M — a +139.3% YoY expansion. The business model shift from crypto HPC hosting (low-margin, spot-priced) to AI data center leasing (contracted, higher-margin) is visible in the gross margin inflection: from negative margins in FY2022 to 45.4% TTM.

Operating losses have widened in absolute dollar terms, but this reflects deliberate build-ahead investment rather than deteriorating unit economics. The Q2 FY2026 (Feb 2026) quarter posted gross profit of $53.8M on revenue of $126.6M — a 42.5% gross margin — the strongest quarterly gross margin in the company's history.

• TTM revenue of $319.3M represents a +121.5% increase vs. FY2025's $144.2M annual figure
• FY2025 gross profit of $42.7M vs. $30.0M in FY2024, a +42.7% improvement despite revenue growing only 5.5%
• Net loss widened to -$231.1M in FY2025 from -$149.3M in FY2024, driven by interest expense on growth-funding debt
• Q2 FY2026 operating income was -$25.9M on $126.6M revenue; operating margin of -20.5%, improving from -35.8% a year prior
• Note: Q4 FY2025 (May 2025) shows a negative revenue figure of -$33.3M — this is likely attributable to a revenue restatement or reclassification event in the filing period and should be treated with caution pending formal 10-K clarification

3b. Balance Sheet Analysis

APLD's balance sheet has undergone a structural transformation, with total assets ballooning from $263.9M in FY2023 to $1.87B in FY2025 — reflecting the aggressive data center build-out program. Total debt grew from $91.8M to $702.9M over the same period, and the $300M bridge facility announced in May 2026 adds further leverage.^[S16]

Net debt stood at approximately $661.3M as of FY2025 (total debt $702.9M less cash $41.6M). With EBITDA still negative on a reported basis, conventional net debt/EBITDA is not meaningful. Liquidity is entirely dependent on capital markets access and contracted lease cash flows, which makes the hyperscaler lease signing critical to credit quality.^[S9]

• Total assets grew +145.1% from FY2024 ($762.9M) to FY2025 ($1.87B), reflecting capitalized data center infrastructure
• Total equity increased to $633.7M in FY2025 from $124.8M in FY2024, driven by equity raises and paid-in capital
• Debt-to-equity ratio of 1.11x as of FY2025 — elevated but manageable given asset-backed nature of the debt
• Cash of only $41.6M at FY2025 year-end leaves minimal operating liquidity buffer absent the bridge facility
• Current ratio and interest coverage data not available from locked data; availability note: specific current asset/liability breakdowns were not provided in the locked fact set

3c. Cash Flow Analysis

APLD's free cash flow profile reflects a company in maximum-investment mode. CapEx accelerated from $131.3M in FY2023 to $681.6M in FY2025, and TTM capex has likely exceeded $1.5B based on quarterly data (Q2 FY2026 alone showed $775.2M in capex). Operating cash flow turned negative in FY2025 at -$115.4M, reversing a brief period of positive operating cash generation in FY2024.

The Q2 FY2026 quarter is notable: operating cash flow of +$55.0M on $126.6M of revenue, a meaningful improvement that suggests stabilized facilities are beginning to generate real cash. However, this was entirely consumed (and more) by $775.2M of capex in the same quarter.

• TTM FCF: -$1.50B — driven almost entirely by construction-phase CapEx, not operating deterioration
• FY2025 capex of $681.6M represents 472% of FY2025 revenue — an extraordinarily high CapEx intensity
• Q2 FY2026 (Feb 2026) OCF turned positive at +$55.0M, the first quarterly positive OCF since FY2024
• FCF per share data is not calculable from locked facts given the pre-profitability profile; this field is omitted

3d. Return on Capital

Return metrics are meaningfully negative across all periods, consistent with a company in heavy build-out mode. ROE, ROA, and ROIC are not meaningful as positive return benchmarks for a pre-profitability infrastructure developer. The more relevant metric is the implied return on invested capital embedded in contracted lease terms relative to construction cost per megawatt.

• All return metrics are negative and not meaningful (N/M) as traditional profitability measures during the investment phase
• Gross margin improvement from 19.9% (FY2023) to 45.4% (TTM) is the best proxy for improving capital efficiency on commissioned assets
• Management's ability to sign full-campus hyperscaler leases at prevailing rates is the true test of return on invested capital

04 Valuation

4a. Multiples Analysis

APLD's valuation multiples are optically extreme on any near-term earnings basis. This is structurally appropriate for a hyper-growth pre-profitability data center developer, but it demands rigorous scenario analysis on the path to normalized earnings. The company trades at 46.9x EV/Revenue and 1,077x EV/EBITDA — multiples that require no margin of safety relative to the current operating profile, only relative to the contracted revenue trajectory.

We benchmark APLD against three direct competitors: Equinix (EQIX), the dominant global colocation and interconnection REIT; Iron Mountain (IRM), a diversified data center and records management operator; and Switch (SWCH) (prior to its acquisition by DigitalBridge, used as historical comp, est.). Given APLD's growth stage, the most relevant peer comps are forward EV/Revenue and EV/EBITDA multiples on a 2-year forward basis.

• APLD trades at 8.1x price-to-book vs. a typical data center REIT range of 3-7x — reflecting growth premium
• Forward P/E of -64.5x is not meaningful; street expects continued losses through at least FY2027
• EV/Revenue of 46.9x compresses to approximately 21x on FY2027E revenue of $712.5M — still a premium but increasingly justifiable
• 10 analysts covering with zero holds or sells; consensus mean target of $61.10 implies +35.4% upside from current levels
• Needham raised its price target by $3 as recently as May 23, 2026^[S10]

Competitor estimates for EQIX and IRM are as of May 2026 (estimated from public filings); Switch (SWCH) was excluded as it is no longer publicly traded following its 2022 DigitalBridge acquisition. Note: PEG ratio is not applicable for APLD (negative earnings) and has been omitted to avoid misleading comparisons.

4b. Discounted Cash Flow (DCF) Analysis

Our DCF model is built on five explicit assumptions that define the investment thesis. All projections are estimates and subject to material revision based on lease signing velocity, construction timelines, and financing conditions.

Key Assumptions:

• Revenue Growth: FY2026E $414.2M (per consensus); FY2027E $712.5M; thereafter 40% / 25% / 18% / 12% in FY2028-FY2031
• EBITDA Margin Trajectory: -15% FY2026E, +5% FY2027E, +18% FY2028E, +28% FY2029E, +35% FY2030E, +38% FY2031E — reflecting stabilized campus maturation
• CapEx as % of Revenue: Declining from ~370% FY2026E to ~80% FY2028E to ~25% FY2030E as build-out phase matures
• WACC: 12.0% (base case) — reflecting high growth stage, meaningful financial leverage, and execution risk premium
• Terminal Growth Rate: 3.0% — consistent with long-run data center infrastructure growth

4c. Valuation Conclusion

At $45.14, APLD trades at a modest discount to our base-case DCF-implied value of $62.00. The stock is not cheap on any current-period metric, but it is undervalued relative to the contracted revenue trajectory embedded in recently signed hyperscaler leases.^[S9]

• Our $62.00 target is derived 50% from DCF base case and 50% from 20x FY2027E EV/Revenue of $712.5M, yielding blended implied price of ~$62
• Margin of safety at current price is thin: the bear case represents -51.3% downside, requiring investors to be confident in the base-case execution path
• Consensus high target of $97.00 aligns precisely with our bull case; the $40.00 low target reflects the bear scenario
• The stock's 52-week range of $6.68–$48.81 illustrates the velocity of re-rating that has already occurred as hyperscaler deal flow materialized

05 Business Model & Competitive Moat

5a. Business Segments

APLD operates across two primary business segments: AI Data Center Hosting (purpose-built hyperscaler-grade facilities with long-term leases) and Cloud Services / HPC Hosting (legacy high-performance computing facilities initially serving crypto miners, now transitioning to AI workloads). A third nascent segment, Colocation / Managed Services, is emerging as the company develops Polaris Forge campuses.

The AI data center leasing segment is the growth engine, commanding significantly higher gross margins (estimated 50%+) relative to legacy HPC hosting (~15-20%). Revenue mix is shifting decisively toward AI infrastructure, which drove the TTM gross margin expansion to 45.4%.

• AI Data Center Hosting: primary growth driver; long-term hyperscaler leases; Polaris Forge 3 campus lease signed as of May 2026^[S9]
• HPC/Crypto Hosting: legacy segment, declining as % of revenue mix; low-margin, spot-price exposed
• Cloud Services: early-stage; not yet material to consolidated revenue; optionality value for future GPU-as-a-service

Note: APLD does not report segment revenue with the granularity used above; the breakdown is estimated from public disclosures and management commentary (as of May 2026). Earnings transcripts were not available in the evidence pack; these estimates should be refined against the 10-K/10-Q filings directly.

5b. Economic Moat Assessment

APLD's moat is narrow but widening. The company does not possess the scale, brand, or network effects of Equinix or Digital Realty, but it has built meaningful switching cost advantages through long-term lease structures and is rapidly accumulating site-specific advantages (power access, land, fiber) that are difficult to replicate.

• Power and land access in chosen markets (North Dakota, Texas) represents a growing quasi-regulatory barrier; new hyperscale-ready sites are increasingly scarce
• Long-term lease contracts (5-10 years typical in hyperscaler deals) create institutional switching costs — once a hyperscaler co-designs a facility, migration is operationally prohibitive
• First-mover positioning in Midwest data center markets provides a geographic cost advantage vs. coastal operators

Overall Moat Assessment: Narrow. APLD has site-specific and contractual advantages but lacks the scale, geographic diversification, and brand recognition to qualify as a wide-moat operator. The moat can widen meaningfully if management successfully executes 3-4 more campus-scale hyperscaler leases and diversifies its tenant roster beyond its current concentration.

06 Growth Strategy & Future Outlook

6a. Growth Drivers

APLD's growth strategy centers on building, leasing, and eventually operating a portfolio of purpose-built AI data center campuses under long-term contracts with hyperscaler tenants. The strategy is capital-intensive but produces highly visible, contracted revenue streams once signed.

Near-Term Catalysts (0–12 months):

• Polaris Forge 3 campus lease revenue contribution beginning to flow through income statement as facility commissions (expected FY2026-FY2027)^[S9]
• Next earnings date July 29, 2026 — the Q1 FY2027 print will be watched for revenue ramp velocity against FY2027E consensus of $712.5M
• $300M bridge facility deployment into additional campus construction, with potential for announcement of additional hyperscaler lease signings^[S16]
• EPS improvement trajectory: four consecutive beats suggests Q1 FY2027 could deliver another positive surprise against the -$0.18 consensus

Medium-Term Drivers (1–3 years):

• Portfolio expansion: multiple Polaris Forge campuses under development provide revenue runway well into FY2028-FY2029
• Tenant diversification: broadening the hyperscaler customer base beyond the initial anchor tenant reduces concentration risk
• Gross margin normalization: as build-ahead costs are absorbed and facilities stabilize, gross margins should approach 55-60% on a stabilized basis (est.)
• Potential REIT conversion or asset monetization strategies to recycle capital at lower cost

Long-Term Opportunities (3–5+ years):

• Secular demand growth for AI training and inference infrastructure is a multi-decade tailwind; hyperscaler capex budgets continue to expand
• Cloud services / GPU-as-a-service optionality could transform APLD from a pure landlord to a managed infrastructure operator at significantly higher margins
• International expansion into power-rich markets (Nordics, Middle East) if US power constraints tighten

6b. Total Addressable Market (TAM)

The global AI data center infrastructure market is one of the fastest-expanding addressable markets in technology infrastructure. Hyperscaler capex commitments alone are tracking toward $200B+ annually by 2027 (est.), of which data center construction and leasing represents approximately 40-50%.

• Global AI data center TAM (est.): $200B+ by 2027, growing at 25-30% CAGR through 2030 (est.)
• US hyperscaler data center leasing market (est.): $50-70B annually by 2027 — APLD's primary addressable market
• APLD's current market share: sub-1% of the US hyperscaler leasing market — massive runway even at modest share capture
• Achievable market share (5-year view): 1-3% of US hyperscaler leasing = $500M–$2.1B in annual revenue — consistent with our FY2030E revenue model

6c. Competitive Positioning

APLD sits in the challenger tier of the US data center market, competing against much larger and better-capitalized operators including Equinix (market cap ~$75B est.), Digital Realty (~$50B est.), and Iron Mountain (~$25B est.). Its competitive advantages are primarily geographic (power-rich Midwest markets), relational (direct hyperscaler engagement model), and temporal (timing its build-out to coincide with record hyperscaler capex demand).

• Market position: Challenger — APLD does not lead any data center market segment but is gaining share in AI-focused hyperscaler leasing
• Key differentiator vs. incumbents: faster time-to-lease on greenfield campus builds; incumbents are constrained by legacy mixed-use campus structures
• Disruption risk: larger operators (Equinix, Digital Realty) could deploy capital into the same hyperscaler AI niche, compressing margins and increasing competition for power-rich sites

07 Management & Governance

7a. Leadership

APLD's management team has navigated a dramatic business model pivot from crypto-focused HPC hosting to hyperscaler AI infrastructure, a transition that required simultaneous operational execution, capital markets sophistication, and enterprise sales capability. The recent earnings beat streak and hyperscaler lease signings suggest the team has developed the institutional credibility required to sustain the growth model.

Note: Detailed executive biographical data (CEO name, CFO name, tenure, board composition) is not available in the locked fact set or evidence pack. The following reflects publicly available information as of May 2026 (estimated). Investors should verify against the most recent proxy statement.

• CEO: Wes Cummins — Co-founder; background in investment management and technology infrastructure; has led the company through its pivot from digital asset mining infrastructure to AI data center development (as of May 2026, estimated)
• CFO: The company has undergone CFO transitions consistent with its rapid growth phase; investors should monitor for leadership stability in the finance function
• Multiple 8-K filings in April–May 2026 related to Corporate Changes & Voting Matters suggest ongoing governance activity^{[S14][S17][S19]}

7b. Capital Allocation Track Record

Management has deployed capital aggressively into data center infrastructure, growing total assets from $8.5M revenue in FY2022 to a $1.87B asset base by FY2025. The capital allocation strategy is explicitly growth-first with no dividends, no buybacks, and minimal debt paydown — all free resources directed toward new campus construction.

• CapEx: $681.6M in FY2025 alone — the defining capital decision of the current strategy; appropriately front-loaded against signed or anticipated leases
• Financing: Mix of equity issuance and debt; May 2026 $300M bridge facility demonstrates continued access to institutional credit^[S16]
• M&A: No material acquisitions identified in the evidence pack during the review period; organic build is the preferred strategy
• No dividends or buybacks — appropriate for the current growth stage

Capital Allocation Rating: Fair. Management has deployed capital into the correct strategic direction (AI infrastructure) and has maintained capital markets access during a period of intense spend. However, the speed of the build — and the reliance on a single bridge facility for near-term liquidity — suggests capital planning could benefit from more permanent long-term financing structures.

7c. Insider Ownership & Alignment

Specific insider ownership percentages and recent insider transaction data are not available in the locked fact set. Investors should consult the most recent proxy statement (DEF 14A) and Form 4 filings on SEC EDGAR for current figures.

• Multiple 8-K filings related to "Corporate Changes & Voting Matters" in April–May 2026 suggest governance activity that may include equity grants, executive changes, or shareholder vote outcomes^[S14][S17]
• Founding-era management with meaningful ownership stakes historically have incentive structures aligned with long-term equity value creation
• The absence of buybacks is neutral to alignment — management is investing capital rather than returning it, consistent with the growth mandate

08 Risk Analysis

8a. Company-Specific (Idiosyncratic) Risks

APLD's risk profile is dominated by execution, financing, and customer concentration risks inherent to a pre-profitability infrastructure developer operating at maximum build intensity. These risks are manageable but not trivial — each represents a pathway to material downside from current valuation levels.

• Financing Risk: TTM FCF of -$1.50B requires continued capital markets access; any disruption to debt or equity issuance could halt construction
• Customer Concentration: Heavy reliance on one or two hyperscaler tenants means a single lease cancellation could impair 30-50% of projected revenue (est.)
• Construction / Commissioning Execution: First large-scale hyperscaler campus builds carry meaningful technical risk — delays defer revenue while interest expense accrues
• EPS negative for foreseeable future: FY2027E consensus EPS of -$0.70 means continued losses; shareholder patience and capital access are prerequisites

8b. Industry & Macro (Systemic) Risks

• Hyperscaler CapEx Cycle Risk: If major cloud providers reduce or defer data center capex (as seen in 2023 pause-and-restart cycles), new lease signings could slow materially
• Power Cost and Availability: Electricity cost increases or grid capacity constraints in target markets (North Dakota, Texas) would compress margins on existing and planned facilities
• Interest Rate Environment: Elevated interest rates increase the cost of the company's debt-heavy capital structure and reduce the NPV of long-duration lease cash flows

09 Final Recommendation

APLD is a high-conviction, high-risk growth investment in the infrastructure layer of the AI compute buildout — the single most capital-intensive technology investment cycle of the current decade. The company has cleared the most important validation hurdle (institutional hyperscaler as anchor tenant), the gross margin inflection is real and accelerating, and the stock trades at a material discount to the consensus analyst target despite a four-quarter beat streak. To own this stock, investors need to believe that hyperscaler demand for purpose-built AI data center campuses remains robust for the next three to five years — a thesis that the current level of global AI infrastructure investment makes highly credible.

10 Open Questions & Narrative Checkpoints

• Question: Who is the Polaris Forge 3 hyperscaler tenant, and what are the lease terms (duration, rate per MW, escalation)?^[S9] Why it matters: The identity and lease economics are not public; if the tenant is a Tier 1 hyperscaler (Microsoft, Google, AWS, Meta) and the lease is 10+ years at market rates, this alone could justify a significant re-rating. A shorter-duration or below-market lease would be materially less valuable.
• Question: What is the path to permanent refinancing of the $300M bridge facility, and at what cost of capital?^[S16] Why it matters: Bridge facilities carry higher interest rates than permanent debt; the cost of permanent capital directly impacts the IRR on each campus build and could compress or expand the equity return on invested capital.
• Question: Will FY2027 revenue consensus converge upward toward the $896M high estimate or pull back toward the $440M low? (as of May 2026, only 1 analyst contributes to FY2027E) Why it matters: With a single analyst providing the FY2027 estimate, the consensus is statistically meaningless; additional analyst initiations will either confirm or deny the growth trajectory embedded in our base case.
• Question: What is the nature of the multiple 8-K Corporate Changes filings from April–May 2026?^{[S14][S17][S19]} Why it matters: Multiple governance 8-Ks in a two-month window could represent routine proxy-related disclosures or could signal material executive transitions; clarity is needed to assess management continuity risk.
• Question: What is the actual campus-level gross margin on the AI data center leasing segment, segregated from legacy HPC/crypto hosting? (as of May 2026, APLD does not report segment-level margin detail) Why it matters: Our thesis assumes AI campus margins of 50%+; if the blended 45.4% TTM gross margin reflects some legacy drag, the underlying AI lease economics may be even stronger — or weaker — than visible in consolidated financials.
• Question: Will the Q1 FY2027 (May 2026 quarter) earnings on July 29, 2026 continue the four-quarter EPS beat streak against the -$0.115 consensus estimate? Why it matters: A fifth consecutive beat would further validate the operating leverage narrative and likely trigger upward consensus EPS revisions for FY2027; a miss — particularly if driven by revenue shortfall rather than investment costs — would be a significant negative signal.
• Question: What are the power procurement terms and PUE targets for the Polaris Forge campus portfolio? Why it matters: Power cost is the single largest variable in data center operating margins; any deterioration in power purchase agreement terms or grid access reliability would compress margins across the entire portfolio.
• Question: Is APLD pursuing a REIT structure, joint venture monetization, or asset-light operating partnership model to reduce the equity capital burden of ongoing construction? Why it matters: A REIT conversion or hyperscaler co-investment structure could dramatically reduce the FCF burn, lower WACC, and unlock a valuation premium; absence of such a structure means continued dilutive equity and debt issuance.

Disclaimer: This report is produced by Basis Report Research for informational purposes only and does not constitute financial advice, an offer to buy or sell securities, or a solicitation of any investment decision. The information contained herein is based on sources believed to be reliable but is not guaranteed as to accuracy or completeness. Past performance is not indicative of future results. Investors should conduct their own due diligence and consult a qualified financial advisor before making any investment decisions. Basis Report Research may hold positions in securities mentioned in this report. All data as of May 27, 2026 unless otherwise noted.

11 Sources & Data As Of

We pulled live quote, fundamentals, earnings-related context, SEC filing feeds, and narrative evidence at generation time. High-impact claims should be tied to Tier 1 sources where available.

Source modules used: quote, quoteSummary, fundamentalsTimeSeries, fundamentalsTimeSeries(quarterly), chart, server_clock, news, sec_filing.

Report Data Retrieval Timestamp: May 27, 2026, 12:38 AM