Applied Digital Corporation (NASDAQ: APLD) designs, builds, and operates next-generation digital infrastructure purpose-built for high-performance computing (HPC) and AI workloads. Headquartered in Dallas, TX, the company IPO'd in April 2022 under CEO Wesley Cummins (formerly Applied Blockchain). The pivot from crypto-hosting toward AI/HPC data centers began in earnest in 2023 and has now crystallized into a hyperscaler-focused "AI factory" strategy.
APLD operates three segments: Data Center Hosting (large-scale facilities leased to hyperscalers on long-term take-or-pay contracts), HPC Hosting (GPU compute clusters for AI training and inference), and legacy Cloud Services (being wound down). The company's campus strategy centers on purpose-built facilities at Ellendale (ND), Garden City (TX), Jamestown (ND), and the newly contracted Polaris Forge 3 (ND), its fourth campus.
On May 20, 2026, APLD announced a landmark 15-year, $7.5B take-or-pay lease with a U.S.-based investment-grade hyperscaler at Polaris Forge 3 (300 MW of critical IT load). This pushed total contracted capacity above 1 GW and total contracted revenue backlog to $31B, with 65% of revenue tied to investment-grade clients. Fiscal year ends May 31 (FY2025 = June 2024–May 2025). Reports in USD.
Investment Thesis
APLD is a high-conviction AI infrastructure buildout play with a now-visible contracted revenue path. The $31B backlog (144x FY2025 revenue) and 1+ GW of contracted IT load provide multi-year earnings visibility that most early-stage AI names lack. The company has successfully completed three hyperscaler leases in 18 months, demonstrating repeatable execution.
Bull drivers: $7.5B Polaris Forge 3 deal (May 20) confirmed hyperscaler demand is real and recurring. Needham analyst John Todaro cited "exceptional visibility and predictability." Direct-to-chip cooling positions APLD to capture next-gen GPU rack densities (Blackwell, GB300). 65% investment-grade counterparty revenue reduces credit risk. Revenue backlog now $31B vs $13.1B market cap — the stock trades at a discount to backlog.
Key risks: FY2025 revenue was only $215.5M — backlog conversion requires massive capex execution over 10-15 years. Stock has 7x'd off lows and trades at 63.5x FY2025 EV/Sales — an extreme premium requiring flawless execution. FCF deeply negative (-$797M FY2025); capex spending ($681.6M) is 316% of revenue. Beta of 5.7 means violent moves are routine. Director-level insider selling in April 2026 at $32-35/share while stock now trades at $45.87.
| Metric | FY2021 | FY2022 | FY2023 | FY2024 | FY2025 |
|---|---|---|---|---|---|
| Revenue | $0.0M | $8.5M | $55.4M | $136.6M | $215.5M |
| Revenue Growth | — | N/M | +548% | +146.6% | +57.8% |
| Gross Profit | -$0.0M | -$13.3M | $11.0M | $30.0M | $22.7M |
| Gross Margin | N/M | -156% | 19.9% | 21.9% | 10.5% |
| Operating Income | -$0.3M | -$20.9M | -$44.1M | -$15.1M | -$72.3M |
| Operating Margin | N/M | -244% | -79.5% | -11.0% | -33.5% |
| EBITDA | -$0.3M | -$20.7M | -$36.9M | -$34.2M | -$91.8M |
| Net Income | -$0.6M | -$23.5M | -$44.6M | -$149.3M | -$231.1M |
| EPS (Diluted) | -$0.06 | -$0.41 | -$0.48 | -$1.31 | -$1.16 |
| Stock-Based Comp | $0.0M | $12.3M | $32.1M | $17.4M | $22.7M |
| D&A | $0.0M | $1.1M | $7.3M | $21.5M | $83.6M |
| Metric | FY2021 | FY2022 | FY2023 | FY2024 | FY2025 |
|---|---|---|---|---|---|
| Cash & Equivalents | $11.8M | $38.8M | $29.0M | $3.3M | $113.9M |
| Total Assets | $15.1M | $120.0M | $264.0M | $762.9M | $1,870.1M |
| PP&E (Net) | $3.3M | $70.7M | $211.2M | $339.4M | $1,294.6M |
| Total Debt | $2.1M | $13.5M | $91.8M | $135.7M | $702.9M |
| Net Debt | -$9.6M | -$25.3M | $62.8M | $132.4M | $589.0M |
| Preferred Stock | $5.2M | $0M | $0M | $0M | $136.0M |
| Stockholders' Equity | -$2.6M | $72.3M | $59.5M | $124.8M | $633.7M |
| Current Ratio | 4.70x | 1.39x | 0.40x | 0.73x | 0.77x |
| Metric | FY2021 | FY2022 | FY2023 | FY2024 | FY2025 |
|---|---|---|---|---|---|
| Operating Cash Flow | -$0.1M | -$0.9M | $58.7M | $13.8M | -$115.4M |
| OCF Margin | N/M | -10.2% | +106.0% | +10.1% | -53.5% |
| Capital Expenditures | -$3.3M | -$55.0M | -$131.3M | -$141.8M | -$681.6M |
| CapEx % of Rev | N/M | 644% | 237% | 103.8% | 316.4% |
| Free Cash Flow | -$3.4M | -$55.8M | -$72.5M | -$128.0M | -$797.0M |
| Net Debt Issuance | $0M | $7.1M | $68.8M | $17.4M | $694.0M |
| Net Stock Issuance | $16.5M | $74.5M | $0M | $130.8M | $358.5M |
| Preferred Dividends | $0M | $0M | $0M | $0M | -$2.6M |
| Multiple | FY2021 | FY2022 | FY2023 | FY2024 | FY2025 |
|---|---|---|---|---|---|
| P/E Ratio | N/M | N/M | N/M | N/M | N/M |
| P/S Ratio | N/M | 32.3x | 14.2x | 3.5x | 60.8x |
| P/B Ratio | N/M | 3.8x | 13.2x | 3.9x | 20.7x |
| P/FCF Ratio | N/M | N/M | N/M | N/M | N/M (FCF -) |
| EV/EBITDA | N/M | N/M | N/M | N/M | N/M (EBITDA -) |
| EV/Sales | N/M | 29.3x | 15.3x | 4.5x | 63.5x |
| Dividend Yield | 0.00% | 0.00% | 0.00% | 0.00% | 0.00% |
| Metric | FY2021 | FY2022 | FY2023 | FY2024 | FY2025 |
|---|---|---|---|---|---|
| Return on Equity | N/M | -29.7% | -64.1% | -119.6% | -36.5% |
| Return on Assets | -3.8% | -19.6% | -16.9% | -19.6% | -12.4% |
| ROIC | -2.3% | -22.4% | -26.8% | -5.8% | -5.4% |
| Asset Turnover | N/M | 0.07x | 0.21x | 0.18x | 0.12x |
| Gross Margin | N/M | -156% | 19.9% | 21.9% | 10.5% |
| Operating Margin | N/M | -244% | -79.5% | -11.0% | -33.5% |
| Interest Coverage | N/M | -186.6x | -22.3x | -0.8x | -4.0x |
| Metric | FY2025A | FY2026E | FY2027E | FY2028E | FY2029E | FY2030E |
|---|---|---|---|---|---|---|
| Revenue (Avg) | $215.5M | $408M | $708M | $1,622M | $2,373M | $4,727M |
| Rev Growth | +57.8% | +89.3% | +73.6% | +129.1% | +46.3% | +99.2% |
| EPS (Avg) | -$1.16 | -$0.61 | -$0.45 | +$0.38 | +$3.91 | +$0.37 |
| # Analysts (Rev) | 6 | 6 | 4 | 3 | 1 | 2 |
| Fwd P/E | N/M | N/M | N/M | 120.7x | 11.7x | 124.0x |
| Name | Title | Type | Shares | Price | Date |
|---|---|---|---|---|---|
| Nottenburg Richard N | Director | Sale | 12,500 | $34.59 | Apr 30, 2026 |
| Nottenburg Richard N | Director | Sale | 12,500 | $32.00 | Apr 28, 2026 |
| Miller Douglas S | Director | Sale | 10,000 | $34.98 | Apr 27, 2026 |
| Cummins Wes | CEO; Chairman | RSU Vest | 100,000 | — | Apr 14, 2026 |
| Cummins Wes | CEO; Chairman | Tax Withhold | 39,350 | $26.26 | Apr 14, 2026 |
| Mohmand Mohammad S | CFO | RSU Vest | 81,667 | — | Apr 7, 2026 |
| Mohmand Mohammad S | CFO | Tax Withhold | 32,136 | $24.56 | Apr 7, 2026 |
$31B contracted backlog is 144x FY2025 revenue — and still growing. The May 20 Polaris Forge 3 deal ($7.5B, 15-year take-or-pay) is the fourth hyperscaler lease in 18 months. Every deal has been with investment-grade U.S. counterparties. At $31B over 15 years, run-rate revenues of $2B+ per year are contractually committed — before any new deals.
1 GW contracted milestone changes the narrative. APLD crossed 1 GW of total contracted IT load with Polaris Forge 3. At industry benchmarks of ~$60-80M ARR per 100MW, 1 GW = $600M-$800M ARR at full occupancy — approaching the FY2027 consensus ($708M). The backlog math is now visible even to skeptics.
Direct-to-chip cooling is a genuine technical differentiator. Zacks (May 21) noted APLD's cooling technology positions it for next-gen GPU densities that air-cooling cannot handle. Blackwell/GB300 racks run at 100-130kW; APLD's liquid cooling gives it preferential access to hyperscaler GPU upgrade cycles.
65% investment-grade client revenue removes credit risk. Unlike neocloud GPU-as-a-service peers (CoreWeave, Nebius), APLD's primary revenue is hyperscaler lease income — more like a data center REIT than a startup. The counterparty quality justifies a structural premium to GPU-rental peers.
Needham reiterated Buy post-Polaris deal. Analyst John Todaro cited "exceptional visibility and predictability into future revenue" as a direct result of the hyperscaler commitment. Citizens also maintained Market Outperform. No downgrades post-announcement.
The valuation prices in flawless $31B backlog conversion over 15 years. At $13.1B market cap vs $31B backlog, the stock prices at 42 cents of backlog NPV per $1 of contracted revenue — implying a ~12% IRR on the backlog after construction costs, which is optimistic if WACC is 15%+.
FCF was -$797M in FY2025 and CapEx will stay elevated through FY2028. Building out 1+ GW of data center campus to satisfy the contracted backlog requires billions more in construction capex — funded by debt and equity raises that the current stock price is implicitly assuming. Each equity raise dilutes existing holders.
Revenue recognition is back-end loaded. Take-or-pay leases don't pay until the facility is complete and occupied. FY2026 consensus of $408M is only +89% growth despite $31B of contracts — because the campuses are still being built. The $31B backlog converts over 10-15 years, not 1-2 years. Investors who assume backlog = near-term revenue will be disappointed.
Director-level insider selling at $32–35. Two board directors sold shares in the open market in late April 2026 at prices 28-30% below today's $45.87. The stock had already run from $6.68 lows — these weren't distressed sells. Board members reducing exposure at $32-35 while the stock runs to $46+ creates an information asymmetry concern.
Construction execution risk is real. APLD is a 150-person company building gigawatt-scale data center campuses. Any permitting delays, power grid connection issues, or contractor disputes could delay backlog conversion. The Ellendale North Dakota location is remote — grid interconnection and fiber access are non-trivial.
Beta of 5.7 means brutal drawdowns are routine. The stock fell from near $48 to $45 in a single session (May 22-23) on "digestion" of the Polaris Forge deal. At beta 5.7, a 10% market correction would historically translate to a 40-57% APLD drawdown.
APLD is building gigawatt-scale campuses in North Dakota and Texas with ~150 employees. Any permitting delays, power interconnection issues, or contractor disputes could delay backlog conversion and push out revenue recognition. Polaris Forge 3 (300MW, North Dakota) requires significant power infrastructure buildout.
FCF was -$797M in FY2025; capex will exceed $1B annually through FY2028 to satisfy the backlog. Every dollar of construction is funded by new debt or equity. A credit market seizure, rating downgrade, or equity market correction that makes dilutive issuance expensive would materially stress the buildout timeline.
EV/Sales of 63.5x on FY2025 trailing revenue assumes 15+ years of backlog conversion at full economics. Any haircut to contract renewal rates, counterparty credit, or pricing at lease expiration compresses the intrinsic value. Analyst PT of $64.40 implies only +40% vs the market's implied perpetuity assumption.
5Y beta of 5.7 means a 10% market correction historically correlates to a 40-57% APLD drawdown. The stock is at 94% of its 52-week range. Single-day moves of ±10% are common. Position sizing and stop-loss discipline are critical — APLD is not a "set and forget" holding.
65% of revenue is investment-grade hyperscaler clients — which is a credit positive but a concentration risk. APLD has 4 campuses with likely 2-3 customers. Non-renewal of a single lease at expiration, or hyperscaler capex pullback (as happened industry-wide in 2022-23), could leave stranded assets. The 15-year durations reduce near-term risk but create cliff exposure in the 2035-2040 timeframe.
$136M of preferred stock was issued in FY2025 with preferred dividends accruing. Preferred has liquidation preference ahead of common. Future financing rounds may include additional preferred tranches. The preferred structure complicates the equity value and creates a class of investors with superior rights. FY2025 preferred dividends of $2.6M will grow if additional preferred is issued.
Polaris Forge 3 construction begins ahead of schedule. Fifth campus announced by end-2026, pushing total contracted capacity to 1.5+ GW. FY2026 revenue tracks toward $420M+ (consensus high). Market rerate to 3-4x EV/FY2027 backlog-derived revenue run-rate. Hyperscaler demand continues to outstrip supply; GPU pricing stays elevated. Fwd EV/Sales compresses to ~20x on FY2027 revenue of $708M+ = ~$14-15B EV = ~$80/share.
Tracks toward last-month analyst consensus PT of $64.40. FY2026 revenue at/near consensus $408M. Ellendale campus begins generating lease revenue. Polaris Forge 3 construction progresses on schedule. No additional equity dilution beyond current preferred structure. Stock holds ~25-30x EV/FY2027 revenue. Modest multiple compression as the run-off-lows momentum fades.
Construction delays push FY2026 revenue below $350M consensus low. Capital markets stress forces dilutive equity round at sub-$40. Hyperscaler AI capex broadly slows (as in 2022-23), reducing urgency of new data center contracts. Market degrades from backlog-NPV pricing to trailing revenue pricing (5x EV/Sales on $408M = $2B EV → ~$20-22/share). Beta of 5.7 amplifies market-wide selloffs. The 200-DMA at $28.57 provides technical support; $20-22 represents a 200-DMA overshoot scenario.
This report was generated using FMP financial data as of May 23, 2026. This is for informational purposes only and does not constitute investment advice. Past performance is not indicative of future results.