Applied Digital (APLD) Q3 2026 Earnings Transcript

Applied Digital reported Q3 revenue of $126.6M (+139% QoQ) and adjusted EBITDA of $44.1M, driven by a full quarter of lease revenue from its 100MW HPC facility, but recorded a GAAP net loss of $100.9M (−$0.60/sh) primarily due to a $59.7M non-cash Cloud write-down and a $52.2M Cloud segment operating loss. The company ended the quarter with $2.1B cash vs $2.7B debt, completed a $2.15B issuance of 6.75% senior secured notes due 2031 for Polaris Forge 2, and achieved an A3 rating on the CoreWeave SPV after lease amendments. Management cites ~$16B of contracted lease revenue (≈$11B CoreWeave / $5.5B hyperscaler), ~1GW of marketed development pipeline, a Base Electron 1.2GW power initiative (Applied to receive ~10% equity), and a target to surpass $1B NOI within five years while maintaining 5–6x NOI leverage.

Analysis

The financing makeover around the CoreWeave SPV is the operational lever that matters most for Applied’s next re-rating: it materially narrows refinancing risk and creates a visible path to lower spread debt for remaining builds. That said, the economic benefit accrues disproportionately to the tenant/DTO structure — meaning the market should treat tenant credit upgrades and landlord execution as separable value drivers rather than a single consolidated one.

The company’s decision to externalize power generation via a minority equity in an IPP is sensible from a downside-protection perspective, but it creates a multi-year optionality layer rather than immediate capacity insurance. If permitting, transmission build or project finance timelines slip, regional grid tightness will intensify and force higher priced short-term sourcing or delay tenant energizations — a scenario that boosts long-run rents but compresses near-term NOI and raises working capital needs.

Strategically, spinning and consolidating Cloud/accelerated compute into a separately capitalized vehicle should unlock targeted capital for GPU fleets while allowing the landlord to focus on long-duration annuities. The second-order winners are GPU providers and DTOs that win OEM and software partnerships (positive for Nvidia and CoreWeave), while other landlord peers without secured long-term investment-grade offtakes will face higher marginal financing costs and slower pipeline monetization.

Key Decisions for Investors

• Long CRWV equity (or a 6–12 month call spread): entry now to capture re-rating from SPV investment-grade recognition and improving DTO economics. Time horizon 6–12 months; target return 30–50% if market continues to reward DTO-led growth. Tail risk: a rapid GPU demand reversal or execution miss could halve short-term gains—limit position size to a single-digit percent of risk budget.
• Relative-value pair — short APLD equity / long CRWV equity (market-neutral sizing): express conviction that tenant credit improvement will outpace landlord equity revaluation due to construction, SG&A, and power-timing risks. Horizon 3–9 months; target 20–35% relative outperformance. Hard stop: 12–15% absolute adverse move on either leg.
• Directional AI hardware exposure — long NVDA Jan-2027 call spread (OTM): levered way to benefit from sustained GPU pricing and rising rental rates that underpin future data center demand. Time horizon 9–12 months; max loss limited to premium, upside levered to continued AI capex expansion. Risk: macro selloff or regulatory headwinds to semiconductor multiples.