Applied Digital Won't Collect Most of Its Rent Until 2029. I'd Wait to Buy the Stock.

Applied Digital has $36 billion of 15-year take-or-pay AI data-center leases covering roughly 1,410 MW, but less than $2 billion, or about 5%, of contracted rent is due before fiscal 2029 because most capacity remains under construction. Fiscal 2026 revenue rose 167% to $611.3 million, yet the company posted a $249.2 million net loss and generated only $99.8 million in AI-campus base rent. Construction spending reached $2.87 billion, while debt totaled $5.0 billion against $4.2 billion of largely restricted cash; senior secured debt costs 6.75%-9.25% and preferred investors rank ahead of common equity. The investment case depends on on-time campus delivery, with delays potentially affecting lease commencements and, in some cases, tenant termination rights.

Analysis

APLD equity is effectively a residual claim on construction execution rather than a conventional contracted-infrastructure cash-flow stream. The relevant underwriting gap is between capex and debt-service requirements incurred during the build phase versus rental cash flow recognized only after commissioning; a modest delivery slip can therefore produce disproportionate dilution, incremental high-cost financing, or a covenant/liquidity event even if the ultimate lease economics remain intact. Restricted cash and preferred capital materially reduce the headline cash balance’s protection for common equity.

CRWV is a second-order exposure: its obligations support APLD's visibility, but its own financing access and GPU-utilization economics determine how credible the tenant-credit premise remains through a weaker AI demand or capital-markets cycle. A slowdown in hyperscaler AI capex would not necessarily impair completed, scarce powered capacity, but it would raise the cost of funding unfinished campuses and reduce APLD's negotiating leverage on expansions. The market is likely underpricing this distinction between demand risk on operating assets and financing risk on assets under construction.

Near term, the stock can rally on individual energization milestones because each delivery de-risks a large block of future rent and unlocks financing capacity. Over the next 1-3 months, however, investors should focus on construction-cost variance, unrestricted liquidity, debt maturities, interest expense, and whether new capital is structurally senior to common—not backlog headlines. Over 6-18 months, on-time delivery could justify a material rerating, but only if incremental funding does not absorb most of the asset-level cash flow before it reaches equity holders.

Key Decisions for Investors

• Remain underweight APLD until the first Polaris Forge 2 delivery is independently commissioned and management discloses unrestricted cash, remaining project capex, and committed—not merely potential—financing. Reassess after the next earnings release; an on-time milestone plus no material increase in senior debt/preferred claims is the key thesis confirmation.
• For a tactical bearish expression, consider a 3-6 month APLD put spread only after a delivery-date slip, capex-overrun disclosure, or financing announcement that adds senior claims. Define risk to premium; avoid outright shorting ahead of milestones given high headline-driven squeeze risk and potentially limited borrow.
• Use CRWV credit and equity performance as an early-warning dashboard for APLD. A sustained widening in CRWV financing spreads, a GPU-utilization/guidance cut, or reduced AI infrastructure commitments would weaken the value of tenant-backed backlog before APLD's reported revenue shows it.
• If execution evidence improves, prefer a staged APLD long rather than buying the backlog narrative today: initiate only after a completed delivery, add after lease-rent conversion is visible in reported operating cash flow, and exit if construction guidance slips by more than one quarter or financing costs rise materially above current project debt levels.