Applied Digital: $16B Contracted, $2.6B in Debt, One Building Complete
600MW under contract with hyperscalers, 100MW operational, 9.25% cost of debt. Six lenses converge on the same conclusion: everything depends on whether APLD can actually build what it has promised.
600MW across 2 hyperscalers, 15-year terms
1 of 6+ planned buildings complete
9.25% senior secured notes, due 2030
First partial quarter, ASC 842 basis
Applied Digital landed $16 billion in contracted hyperscale lease revenue. The stock market rewarded that with a $7.2 billion market cap. On paper, this is one of the clearest AI infrastructure stories in the public markets: build data centers, lease them to hyperscalers, collect rent for 15 years.
The reality is more nuanced. When we ran APLD through our 6-lens multi-model analysis, every single lens converged on the same variable: construction execution. The contracts are genuinely well-structured. The demand is structural. The location advantages are real. But the company has completed exactly one 100-megawatt building and is attempting to scale to 5 gigawatts across multiple simultaneous campuses, a 50x increase in operational complexity.
Meanwhile, Q2 FY26 revenue of $126.6 million (up 250% year-over-year) includes $73 million in one-time tenant fit-out services at mid-single-digit margins. Strip that away, and the AI data center lease business generated $12 million in its first partial quarter. The gap between narrative and operational reality is measured in years, not quarters.
Want the full 6-lens analysis with signal assessments and model debates?
Opus + Sonnet ensemble. 6 lenses. 9 signals. 8 debates. Full evidence citations.
Signal Assessments: What the Committee Found
Revenue headline distorted by $73M one-time fit-out. Actual recurring lease revenue: $12M. $40M+ SBC in 6 months.
15-year contracts with make-whole provisions are strong, but revenue requires completing construction. Only 100MW of 600MW is operational.
$2.6B debt at 9.25% against minimal recurring revenue. ~$217M annualized interest. Must refinance by 2030.
Macquarie partnership is innovative. But 9.25% cost of debt and simultaneous multi-GW construction create execution risk.
Dakota advantages (cheap energy, cold climate, PUE 1.18) are real but replicable. Qualified with 5 of 6 hyperscalers.
Market prices $16B thesis. Operations show $12M quarterly lease revenue and one completed building. 3-5 year gap.
$7.2B market cap on ~$48M annualized lease revenue implies substantial execution premium.
C-suite net positive via grants. Directors net selling. COO reducing exposure. No performance-based comp metrics.
Supportive regulatory environment in North Dakota. South Dakota sales tax exemption is only gating regulatory item.
Key Findings
Revenue Is 58% One-Time Construction Services
Of Q2 FY26's $126.6M revenue, $73M came from tenant fit-out services for CoreWeave at mid-single-digit margins. The legacy bitcoin hosting business contributed $41.6M. The AI data center lease business generated $12M in its first partial quarter. The 250% year-over-year growth headline accurately reflects total revenue but masks the composition.
9.25% Debt Creates a Race Against Time
The $2.35B in senior secured notes due 2030 carry ~$217M in annual interest expense, more than the company's total current revenue. Management describes this as development-phase financing with plans to refinance once buildings are operational. A 12-18 month construction delay would compress the refinancing window uncomfortably.
Contract Structure Is Genuinely Strong
Noncancelable 15-year leases with 100% make-whole provisions. If a customer cancels, they owe the full 15 years of payments. Transferability restricted to equal or higher credit ratings. These are among the strongest protections in infrastructure leasing.
Two-Customer Concentration Is Binary Risk
CoreWeave holds the $11B anchor contract but is not investment-grade. The second tenant is investment-grade but unnamed. A third customer from the 900MW pipeline would materially de-risk concentration.
Where Models Disagreed
Is $2.6B Debt Appropriate or Overleveraged?
Stress Scanner View
Standard infrastructure project finance sized against contracted future cash flows. Macquarie structure is sophisticated and appropriate.
Fugazi Filter View
The 9.25% coupon reflects lenders pricing meaningful construction risk. More concerning than typical project finance given pre-revenue status.
Resolution: Converged on STRETCHED. Appropriate structure, but a 12-18 month delay would convert STRETCHED to STRAINED.
Is the Competitive Moat Defensible or Temporary?
Temporal View
Cheap land and energy in the Dakotas are available to anyone. Location advantages are real but not proprietary.
Compounding View
Hyperscaler qualifications, proven construction capability, and locked supply chain create a multi-year lead that compounds with each completed building.
Resolution: CONTESTED with a window. If APLD reaches 5GW+ before competitors replicate, the position could become DEFENSIBLE.
Is the CEO's Net-Positive Position Genuine Confidence?
Wes Cummins received 1.5M shares in RSU grants and sold 1.07M shares. Is this conviction or grant-funded appearances?
Resolution: Conviction with monetization. Genuine belief in the long-term thesis combined with rational diversification. Not alarming, but not bullish either.
Cross-Lens Reinforcements
Construction Execution Is the Binding Constraint
All 6 lenses converge: revenue durability, funding adequacy, competitive position, and narrative credibility all depend on building multi-GW campuses on time and on budget.
Contract Protections Are Best-in-Class
Gravy Gauge and Stress Scanner agree: 15-year noncancelable terms with 100% make-whole and restricted transferability provide genuine structural protection once buildings are complete.
Revenue Presentation Requires Careful Parsing
Fugazi Filter and Myth Meter agree: the 250% growth headline and $16B prospective revenue framing overstate the current state of the business.
What to Watch
Any 3+ month delay on Polaris Forge 1 Buildings 2-3 or Polaris Forge 2 would trigger re-assessment of funding adequacy and narrative sustainability.
CoreWeave holds the $11B anchor contract and is not investment-grade. Monitor quarterly filings for credit deterioration.
A signed lease with a third hyperscaler by mid-2026 would materially de-risk customer concentration.
Successful refinancing at lower rates validates the construction thesis. No progress by 2028 would signal funding concerns.
HIGHER SCRUTINY
Applied Digital presents a compelling but early-stage infrastructure thesis. The structural demand for AI data centers is real, the contracts are well-structured with best-in-class protections, and management has demonstrated initial execution capability. However, the gap between narrative ($16B contracted, $1B NOI target) and current operations ($12M quarterly lease revenue) is substantial. The $7.2B market cap prices in multi-year execution that has not yet occurred.
Path to More Favorable Assessment
- • Polaris Forge 1 Buildings 2-3 delivered on schedule
- • Third hyperscaler lease signed and announced
- • Refinancing of 9.25% notes at substantially lower rates
- • CoreWeave achieves investment-grade credit rating
Path to Less Favorable Assessment
- • Construction delays of 3+ months on any building
- • CoreWeave financial deterioration or contract renegotiation
- • Failure to sign third customer by mid-2026
- • Insider selling acceleration, particularly CEO or COO
This analysis is for educational purposes only — it is not a recommendation to buy or sell any security.
Public Sources Used (15 documents)
- Annual Report (10-K) — FY2025
- Quarterly Report (10-Q) — Q2 FY2026 (Nov 2025)
- Quarterly Report (10-Q) — Q1 FY2026 (Aug 2025)
- Quarterly Report (10-Q) — Q4 FY2025 (Feb 2025)
- Quarterly Report (10-Q) — Q3 FY2025 (Nov 2024)
- Current Reports (8-K) — 10 filings
- Schedule 13D/A — 3 filings
- Schedule 13G — 3 filings
- Form 4 — 20 insider transaction filings
- Form 144 — 10 proposed sale filings
- Q2 FY2026 Earnings Call Transcript
- Q1 FY2026 Earnings Call Transcript
- Q4 FY2025 Earnings Call Transcript
- Q3 FY2025 Earnings Call Transcript
- CourtListener Litigation Search
Full Analysis with Signal Breakdowns
Explore the complete 6-lens assessment including debate transcripts, evidence citations, and monitoring triggers.
View APLD Analysis