Applied Digital Corp. (APLD) Q4 2025 Earnings Summary
Executive Summary
- Q4 FY25 revenue of $38.0M grew 41% YoY and modestly beat S&P Global consensus ($38.01M actual vs $37.12M estimate; +$0.9M) as hosting capacity ran at full utilization; Primary EPS (S&P methodology) was -$0.03 vs -$0.17 consensus, a significant beat driven by lower interest expense and improved operating leverage in hosting (consensus values marked with *; see S&P Global note below).
- Strategic step-change: two 15-year CoreWeave leases (250MW) signed during the quarter (~$7B contracted revenue), and an additional 150MW option exercised post-quarter that would bring total to 400MW and ~$11B contracted revenue if finalized; first 100MW building targeted to be operational in calendar Q4’25 .
- Directional guide: management expects a significant sequential revenue increase beginning in the quarter ending Aug 2025 from Polaris Forge 1 “technical fit-out” (customer-funded at a small margin) ahead of lease revenue commencement; project financing expected to close in ~4–10 weeks from the call date .
- Balance sheet and funding: ended Q4 with $120.9M cash and $688.2M debt; subsequently raised ~$268.9M via ATM and Series G preferred, improving liquidity to support build-out .
What Went Well and What Went Wrong
-
What Went Well
- Signed two transformative 15-year leases with CoreWeave (250MW, ~$7B revenue) and noted post-quarter exercise of a 150MW option that would take capacity to 400MW and ~$11B total contracted revenue if finalized: “These long-term leases mark a defining moment for Polaris Forge 1” — CEO Wes Cummins .
- Operational progress and cost advantages: first 100MW liquid-cooled facility on track for 2H’25; design targeting PUE of 1.18 and near-zero water, with management estimating a 100MW customer could save ~$2.7B over 30 years vs other regions .
- Interest expense fell sharply YoY ($4.5M vs $13.8M), aiding EPS outperformance despite higher SG&A; Adjusted EBITDA turned positive to $1.0M .
-
What Went Wrong
- SG&A inflation: $28.1M (+115% YoY), driven by $9.4M higher stock-based comp (accelerated vesting and PSUs), +$3.4M personnel, and +$2.3M other expenses (insurance, software) .
- Core non-GAAP profitability remains thin in the quarter (Adjusted EBITDA $1.0M; Adjusted net loss -$7.6M), reflecting transition-stage cost structure and ramp timing .
- Cloud Services reclassified as discontinued operations; strategic alternatives remain under review (added uncertainty and reporting complexity) .
Financial Results
Q4 FY25 vs S&P Global Consensus
Notes: APLD reported GAAP continuing ops EPS of -$0.12 and Adjusted EPS of -$0.03 for Q4 FY25 .
S&P Global values marked with * (Values retrieved from S&P Global).
Quarterly trend – continuing operations focus (revenues) and key P&L items (as reported each quarter)
Margins and mix
Segment revenue (as disclosed in each period)
KPIs and balance sheet
Guidance Changes
Earnings Call Themes & Trends
Management Commentary
- “These long-term leases mark a defining moment for Polaris Forge 1... designed to scale up to 1 gigawatt” — Wes Cummins, CEO, on CoreWeave agreements .
- “We’ve reduced projected build times from 24 months to 12 to 14 months” via SKU reduction, supplier consolidation, and repeatable design .
- “This design is expected to achieve a PUE of 1.18 and a near-zero water consumption... a 100 MW data center customer could save up to approximately $2.7 billion over a thirty year period” .
- “We’re actively working with our financing partners to finalize the project financing... over the next four to ten weeks” — CFO Saidal Mohmand .
- “With the CoreWeave lease, we believe we’re now roughly halfway toward our internal goal of generating $1 billion in annual net operating income over the next three to five years” — CEO .
Q&A Highlights
- Development cadence: Expect to break ground on one, potentially two, additional campuses before year-end, beyond Polaris Forge 1 .
- Financing timing and gating items: 4–10 weeks; seasonal slowdown and third-party advisors can cause lag; identified lead banking partner aligned to expedite .
- Additional customer pipeline: Advanced negotiations with an investment-grade North American hyperscaler; broader discussions with 4–6 hyperscalers across Dakotas and other regions .
- Ellendale fit-out and ramp: Mostly fit-out underway; customer gear and cabling to follow; ramp expected in calendar Q4 (Oct–Nov) .
- Delivery obligations and model: Standard late-delivery penalties apply; company focused on “full-stack colo” vs powered shell, given public-company economics .
Estimates Context
- S&P Global consensus for Q4 FY25: Revenue $37.12M (8 ests) and Primary EPS -$0.17 (7 ests); reported revenue $38.01M and S&P Primary EPS actual -$0.03 — both beats (values marked *; Values retrieved from S&P Global).
- APLD’s reported GAAP cont. ops EPS was -$0.12; Non-GAAP Adjusted EPS was -$0.03, aligning with S&P’s Primary EPS actual .
- Near-term sell-side models likely need higher revenue for the next two quarters to reflect customer-funded fit-out recognition ahead of lease revenue commencement, per CFO .
Key Takeaways for Investors
- Structural step-change from CoreWeave agreements (250MW signed; 150MW option exercised post-quarter) with potential ~$11B contracted revenue over ~15 years if the option’s third lease is finalized, materially improving multi-year visibility .
- Near-term revenue tailwind from fit-out recognition across the quarters ending Aug and Nov 2025, even before lease revenue starts; watch for margin mix (fit-out at small margin) .
- Financing de-risking likely near term (4–10 weeks), with terms for IG tenants typically implying LTC ~70%; added ~$269M equity raised post-quarter improves runway .
- Operating leverage opportunity as hosting scales and interest expense remains lower YoY; yet SG&A discipline is key given Q4’s elevated stock-based comp .
- Execution catalysts: on-time delivery of the 100MW facility in calendar Q4’25; lease execution on the 150MW option; potential additional hyperscaler leases in pipeline .
- Watch classification/segment reporting (Cloud Services discontinued) and non-GAAP adjustments (debt fair value changes, SBC, diligence/litigation) to assess core profitability trajectory .
- Medium-term thesis: repeatable “AI factory” build model, speed-to-power advantages in the Dakotas (target PUE 1.18; near-zero water), and expanding hyperscaler relationships can compound contracted revenue and NOI over 3–5 years .
S&P Global note: Asterisked values in the “vs Consensus” table are from S&P Global. Values retrieved from S&P Global.