Applied Digital Just Borrowed Another $2 Billion. Its Biggest Customer Is $21 Billion in Debt.

Applied Digital added $2.15B of new debt at a 6.75% interest rate, taking total debt to roughly $5B from $468M a year ago. Its $16B contracted lease pipeline is heavily concentrated: ~$11B (nearly 70%) depends on CoreWeave, which carries over $21B of debt and lost $1.2B last year. CoreWeave’s solvency is tied to OpenAI, which projects $14B losses in 2026 and $115B through 2029 and only has $25B of near-term confirmed funding from a reported $110B round. The position creates significant counterparty and funding risk for Applied Digital investors given rising leverage and reliance on highly indebted AI customers.

Analysis

The critical risk is a concentrated counterparty waterfall: a single large tenant under stress can turn lease revenue into a near-term liquidity crisis for a highly levered data-center owner because lenders accelerate, capex stalls, and equipment salvage values compress. Expect covenant tests and potential liquidity runs to crystallize on a weeks-to-months cadence; if one major customer pauses capacity payments, lenders typically seek cure or collateral within a single quarter and asset recovery timelines stretch into 6–18 months.

Hardware and chip demand faces a two-stage response. In the near term, stressed cloud customers and their financiers will curtail incremental GPU orders, pressuring OEMs and distributors for 1–2 quarters; in the medium term, consolidation concentrates demand into a smaller set of hyperscalers, which benefits component makers with scale (NVIDIA-style pricing power) while penalizing mid-tier integrators and smaller cloud operators.

Credit markets are the transmission mechanism: tightening funding availability and higher all-in yields make refinancing fragile-name specific and can produce >1000bp spread moves for speculative credit if investor sentiment turns. The asymmetric upside for the owner comes from repurposing capacity into diversified colocation or securing government/enterprise anchors — a recovery that would likely take 12–36 months and requires fresh secured capital or an M&A backstop to reprice risk materially lower.

Key Decisions for Investors

• Short APLD equity via put spread: buy 9-month APLD 25% OTM puts and sell 9-month 40% OTM puts to cap cost. Position size 1–2% NAV. Rationale: limited-cost downside if liquidity/covenant stress appears; target asymmetric payoff if equity falls 40–60% within 6–12 months. Stop-loss: close if premium falls >50% from entry or if company announces fully underwritten secured financing extension.
• Pair trade — short CRWV equity and long NVDA call spread: short CRWV outright (or buy 6–9 month 20% OTM puts) sized to 1% NAV and buy NVDA 6–9 month 15–30% call spread sized to 0.6x notional to the short. Rationale: capture idiosyncratic tenant/credit risk while keeping exposure to upside in chip scarcity/consolidation. Exit on either a) CRWV credit spread tightening to pre-distress levels, or b) NVDA materially re-prices (>20% move) or on option expiry.
• Hedge existing infrastructure exposure: buy 12-month APLD (or sector) downside protection — long single-name puts or long VIX-term exposure if you hold related assets. Size to offset 25–50% of notional exposure. This is insurance against a liquidity-triggered equity gap risk that can occur within 30–90 days; cost is acceptable given asymmetric tail risk.
• Event-driven re-entry for risk-seeking capital: set alerts for distressed-asset auctions or announced sale processes (lender-forced marketing). If assets hit secondary market discounts >30% vs replacement cost, initiate selective long positions in secured debt or buy-out equity with staged capital (tranche in at milestones). Time horizon 12–36 months; target IRR >25% conditional on successful re-anchoring or lease diversification.