Dossier · APLD · Watchlist

APLD · Applied Digital Corporation · Stock research

Q: Is APLD a buy?

orbyd does not issue buy or sell ratings — it publishes a dated, falsifiable thesis and the exact trigger that would prove it wrong. Operating story and equity have decoupled: >1 GW contracted and ~$31B backlog, yet the stock halved from $49.65 (2026-05-28) to $25.79 (2026-07-17) on a ~$2.7B debt load, widening GAAP losses and CoreWeave concentration. Price now sits on the $25.56 April base with the 2026-07-27 FY26 print as the binary. Trend broken; stand aside into the print.

LOW Special situation Catalyst · 2026-07-27 gpu-cloud-neoclouds

Last analysed · 2026-07-19T09:07:45+00:00

Current thesis

Operating story and equity have decoupled: >1 GW contracted and ~$31B backlog, yet the stock halved from $49.65 (2026-05-28) to $25.79 (2026-07-17) on a ~$2.7B debt load, widening GAAP losses and CoreWeave concentration. Price now sits on the $25.56 April base with the 2026-07-27 FY26 print as the binary. Trend broken; stand aside into the print.

Invalidation trigger

A weekly close below $25 confirms loss of the post-April-earnings base floor near $25.56 and opens the $20 zone; secondary breaks: an equity raise or ATM reactivation disclosed on or after the 2026-07-27 FY26 call, a Polaris Forge lease cancellation, or a CoreWeave counterparty downgrade.

Thesis status

Invalidated resolved 2026-04-29 published trigger fired How this is scored →

Latest analysis and events for APLD — 2026-04-19

As of 2026-04-19, orbyd's latest analysis for Applied Digital Corporation (APLD): Applied Digital BTC→AI pivot (less proven than IREN but same thesis).

Invalidation trigger: A weekly close below $25 confirms loss of the post-April-earnings base floor near $25.56 and opens the $20 zone; secondary breaks: an equity raise or ATM reactivation disclosed on or after the 2026-07-27 FY26 call, a Polaris Forge lease cancellation, or a CoreWeave counterparty downgrade.

Next dated event on file: 2026-07-27 — catalyst in 8d.

Current Thesis

The operating story and the equity have fully decoupled. Applied Digital keeps landing the things a BTC-miner-turned-AI-landlord is supposed to land >1 GW contracted, ~$31B of baseline lease revenue, 175 MW of Polaris Forge 1 live while the stock has given back roughly half its value from the 2026-05-28 closing high of $49.65 to $25.79 on 2026-07-17. That is not a pullback inside an uptrend; it is a nine-week repricing on volume, and the market is telling you what it now cares about: a ~$2.7B debt load, a GAAP net loss of $100.9M in the March quarter, and a single anchor tenant. The fiscal Q4/FY26 print on 2026-07-27 (after the close, call 5:00 p.m. ET) is the event that either arrests this or confirms it. Until then the tape sits directly on the $25.56 shelf that has held since the April print, with the next visible air pocket down toward $20. Buying into an eight-day earnings window on a name with a beta near 5.7 and a broken trend is a gamble on the print, not an expression of the pivot thesis.

Bullish and bearish views on Applied Digital Corporation

The model's bull view on Applied Digital Corporation (APLD), in brief: Contracted revenue is real and long-dated: total contracted capacity crossed 1 GW in late May on a 15-year lease with an investment-grade U.S. hyperscaler (~$5.2B contract value), lifting baseline contracted lease revenue to ~$31B and up to ~$73B if renewal options are exercised… The bear view: Price is the primary evidence. From 2026-06-12 at $42.70 to $28.84 on 2026-07-13 to $25.79 on 2026-07-17 roughly $4B of market cap removed in a month during which every announced datapoint was positive. Good news absorbed without a bid is distribution. Leverage has become the… Both cases follow in full.

Bull Case

Contracted revenue is real and long-dated: total contracted capacity crossed 1 GW in late May on a 15-year lease with an investment-grade U.S. hyperscaler (~$5.2B contract value), lifting baseline contracted lease revenue to ~$31B and up to ~$73B if renewal options are exercised across ~1.2 GW of IT load.
Execution is landing on schedule: the second Polaris Forge 1 building hit ready-for-service 2026-07-01, taking live capacity to 175 MW of the 400 MW contracted at full build.
2026-06-22: an MDU Resources subsidiary signed a power-supply agreement for Polaris Forge 3 in North Dakota, removing grid interconnect normally the longest pole in a data-center build as the binding constraint on the newest campus.
2026-06-09: $1.59B of 7.00% senior secured notes due 2031 priced, funding the Polaris Forge 1 build with debt instead of the equity ATM that capped the shares through the spring.
Revenue TTM $319.3M, +66.1%, with Q3 FY26 (2026-04-08) at $126.6M, +139% YoY billed hosting and GPU revenue, not signed pipeline.
Sell-side is uniformly constructive: ten analysts at Strong Buy consensus, average target ~$76.70 against a $25.79 quote; Northland reiterated Outperform with a $82 target on 2026-06-17.

Bear Case

Price is the primary evidence. From 2026-06-12 at $42.70 to $28.84 on 2026-07-13 to $25.79 on 2026-07-17 roughly $4B of market cap removed in a month during which every announced datapoint was positive. Good news absorbed without a bid is distribution.
Leverage has become the story: ~$2.7B total debt after the June notes, with a further tranche still to be placed for Polaris Forge 1. At 7.00%, the new paper alone carries north of $110M of annual interest against a company that lost $100.9M ($0.36/sh) in its March quarter.
Customer concentration: CoreWeave anchors Polaris Forge 1 and is itself a neocloud financing its own build. Counterparty stress there transmits directly into APLD's largest contracted campus, and the market has begun pricing that linkage.
Nvidia's exit remains an unhealed overhang it sold all 7,716,050 shares (~$177M) as disclosed in the 13F covering 2025-12-31, and the name has not re-established a strategic-validation anchor since.
Residual BTC beta persists: on 2026-06-24 APLD traded down with the crypto-linked cohort on a Bitcoin drawdown despite the AI revenue mix. A beta of 5.68 means the equity moves on flows that have nothing to do with lease economics.
Positioning is one-sided. Zero Hold or Sell ratings against a stock down ~49% from its high leaves the entire downgrade path unpriced; the first cut lands into an empty book.
2026-07-16/17: televised commentary framed the name as a "start it small, don't start it big" idea on the losses, and the stock still fell more than 6% the following session. Retail-facing endorsement failing to produce a bounce is a demand problem, not a coverage problem.

Setup & Price Structure

Trend is broken on every timeframe that matters. The 2026-05-28 closing high of $49.65 marks the top of the move; price has made lower highs and lower lows continuously since 2026-06-12 and now trades below its 20-, 50- and 100-day averages with the shorter averages rolling over beneath the longer ones. The 52-week range is $9.79–$50.73, so the current $25.79 sits almost exactly mid-range no valuation cushion from the lows, no momentum from the highs. The operative structure is the $25.56 base floor established after the 2026-04-08 print, which price is now testing for the first time from above. Losing it on a weekly basis opens a gap zone with little transactional support until the $20 area, a level already being flagged as the next reference. There is no reclaim setup yet: a first credible one would require a close back above the ~$31 shelf that broke on 2026-07-10, ideally on a post-earnings expansion bar. Adding exposure into this structure at lower prices, on the argument that the backlog got bigger, is exactly the behaviour the chart is punishing.

Catalyst Calendar (next 30 days)

2026-07-27 (confirmed) Fiscal Q4 and full-year 2026 results, released after the close, conference call 5:00 p.m. ET. Fiscal year ended 2026-05-31. This is the binary: FY revenue and, more importantly, the first full guide that incorporates the 7.00% notes' interest burden and the remaining Polaris Forge 1 financing tranche. An eight-day window ahead of a print on a 5.7-beta name is a stand-aside, not an entry.
~2026-07-27 to 2026-08-10 (est.) Post-print financing disclosure. The unplaced Polaris Forge 1 debt tranche is the single most consequential unknown; whether it prices as debt or is supplemented with equity determines the dilution path.
~2026-08-05 to 2026-08-20 (est.) Sell-side target revisions. With ten Strong Buy ratings and an average target near $76.70 against a $25.79 quote, the gap between price and published targets is the widest it has been; the resolution mechanism is target cuts, not price rallies.

Elapsed catalysts

Ongoing Polaris Forge 1 building-three ready-for-service milestone (400 MW contracted, 175 MW live as of 2026-07-01). Any slip announcement is a direct thesis event. _(passed 18d ago)_
Ongoing Bitcoin tape. Crypto drawdown days still drag the name, as on 2026-06-24. _(passed 25d ago)_

What Would Change Our Mind

Bullish re-rate: a fiscal Q4 print on 2026-07-27 that narrows the GAAP loss meaningfully while confirming remaining Polaris Forge 1 financing is placed as debt, followed by a weekly close back above $31. That combination would mark the repricing as complete and the pivot as funded.
Bullish structural: a second named investment-grade hyperscaler lease that reduces CoreWeave concentration below a majority of contracted MW, or a new strategic equity holder replacing the Nvidia validation lost in the December-quarter 13F.
Bearish confirmation: losing the $25.56 April base on a weekly closing basis, which removes the last defended shelf above the $20 zone.
Bearish structural: an equity raise or ATM reactivation disclosed on or after the 2026-07-27 call that would signal the 7.00% notes cannot carry the build and reset the share count against a backlog that does not convert to cash for years. Equally: a Polaris Forge lease cancellation, a delivery-milestone slip, or a CoreWeave funding stumble.
Neutral-to-negative: the print passes without incident and price still fails to reclaim $31 within two weeks. A stock that cannot rally on a clean catalyst has a supply problem no news flow fixes.

Correlation Notes

Trades as the higher-beta, less-proven expression of the same BTC-miner-to-AI-landlord trade as IREN, whose pivot is further along and whose tape has held up better. The spread between the two is the cleanest read on whether the market is discriminating on execution or selling the cohort wholesale.
Cohort peers: CIFR, CORZ, WULF, GLXY on the miner-pivot side; CRWV directly on the counterparty side. Neocloud group momentum cooled from accelerating to maturing in early June and the group tailwind that carried the complex through spring is gone.
Residual correlation to Bitcoin remains material despite the AI revenue mix confirmed again 2026-06-24. On crypto-driven down days the move is flow, not fundamentals; on debt- or customer-driven down days it is repricing. Distinguishing the two is the whole job here.
Rate-sensitive by construction. A ~$2.7B debt stack with an unplaced tranche makes the name a levered proxy on credit conditions for AI infrastructure, which is why it now trades more like a levered developer than like a semiconductor beneficiary.
Theme state: the underlying AI data-center buildout is still accelerating, but this specific equity has moved into a maturing-to-contested phase where the narrative no longer sets the price. Fundamentals compounding while the multiple compresses is the signature of a story the market has finished paying up for.

Notes

2026-04-19: Applied Digital BTC→AI pivot (less proven than IREN but same thesis)
No earnings blackout currently Q3 printed 2026-04-08
next print (Q4/FY-end) expected mid-July 2026 (fiscal year ends May).
Archetype changed 7→4 (Legacy Pivot): the pivot is literally underway with a revenue inflection already printed, not emergent speculation.
Pair-trade candidate vs IREN (proven pivot) APLD is the higher-beta expression of the same trade.
Hard rule: do NOT average down below 2026-04-09 fade low. Re-entry only on clean breakout above 2026-04-08 post-print high with catalyst.
Next dossier refresh must pin exact 2026-04-08 high
2026-04-09 low
20-EMA and 50-EMA levels for machine-checkable invalidation.
BTC beta still material despite AI revenue mix tourist money confuses the cohort on BTC-volatile days.
Archetype: Legacy Pivot revenue inflection already printed 2026-04-08
not emergent speculation.
Pair-trade vs IREN: APLD is the higher-beta, less-proven expression of the same BTC→AI thesis.
Hard rule: NEVER average down below 2026-04-09 fade low. Re-entry only on clean breakout above 2026-04-08 post-print high on catalyst news.
No earnings blackout currently Q3 printed 2026-04-08
next print (Q4/FY-end
fiscal year ends May) expected mid-July 2026.
BTC beta still material despite AI revenue mix cohort gets confused on BTC-volatile days
use that as entry window not exit panic.
Next dossier refresh MUST pin exact 2026-04-08 post-print high
2026-04-09 fade low
20-EMA and 50-EMA levels for machine-checkable invalidation.
Catalyst placeholder 2026-05-15 = earliest plausible Ellendale anchor 8-K window; no confirmed event on calendar
re-check weekly.
Cap sizing at 2% (MEDIUM) on any fresh entry higher-beta
debt-funded capex
ATM shelf risk = not SUPREME until anchor lease prints.
Entry conviction HIGH.
Archetype is 4 (Legacy Pivot), NOT 7 the BTC→AI pivot has a printed revenue inflection (Q3 2026-04-08, $126.6M vs $76.6M cons) AND two booked hyperscaler anchor leases (4/23, 5/20). This is no longer emergent speculation. Prior header said 7; correcting to 4.
No earnings blackout currently. Q3 printed 2026-04-08; next print is Q4/FY-end (fiscal year ends May), expected ~mid-July 2026 OUTSIDE the 30-day window from 2026-06-04, so catalyst_date=null.
Lease-disclosure cadence is monthly (4/23 → 5/20). Watch for a third 8-K lease ~mid-to-late June; that is the next narrative-extension trigger but is NOT a confirmed calendar event.
ATM-shelf dilution is the single biggest structural risk. Any equity-raise 8-K. Capex is debt+equity funded.
BTC beta still material despite AI revenue mix tourist money confuses the cohort on BTC-volatile days. Use a BTC-driven flush as an ADD window (on 20-EMA retest), not an exit panic.
Hard rule: NEVER average down below the 2026-04-09 fade low. Re-entry/adds only on a clean higher-low or 20-EMA retest with catalyst.
Saturation watch: 4-analyst PT cluster + repeated 'why is APLD moving' headlines = retail tourism thick. Theme is ACCELERATING but tipping toward MATURING. If it flips SATURATED with PTs matched and no new lease, downgrade and trim.
We are already theme-stacked (HUT/WULF/IREN cohort). Do NOT add correlated GPU-cloud exposure APLD is the higher-beta expression; treat the basket as one risk unit.
Real campus brand is Polaris Forge (prior dossier's 'Delta Forge' was a misnomer — correct it). Forge 1 = CoreWeave 400MW ~$11B, converted to investment-grade SPV; Forge 2 = $5B IG hyperscaler; Forge 3 = $7.5B/15yr take-or-pay, 300MW. Total contracted backlog ~$31B, capacity past 1.2GW (June 2026).
Earnings: Q3 FY26 printed 2026-04-08 (rev $126.64M, +139% YoY, adj EPS positive). Next print Q4/FY-end est. ~2026-08-10 to 08-20 (fiscal year ends May 31) NO blackout until mid-August. Corrects prior 'mid-July' estimate; long clean runway, no near-term binary.
Archetype: Legacy Pivot: BTC-miner→AI-infra-landlord re-rate is the active narrative leg; pivot now largely proven via the ~$31B contracted backlog.
Analyst saturation: 26 Buy / 0 Hold / 0 Sell, avg PT ~$61 (range $40–$106), Compass Point $70 from $45 (2026-05-22). Upgrade-catalyst well is dry; the May PT-cluster acceleration signal is already spent treat as a late-stage crowding flag.
Price reset: ~$42 on 2026-06-05 (range $39.92–$43.75), down ~12% from the ~$48–50 late-May high, into the ~$40–44 20-EMA retest zone. This is the pullback-to-support the late-May extended levels lacked.
Hard rule: NEVER average down below the ~$34 April pivot base. Re-entry only on a clean higher-low / breakout-retest with a catalyst.
ATM/equity-raise 8-K is the dominant structural risk capex is debt+equity funded. Any raise disclosure is an immediate de-risk trigger regardless of price.
BTC beta still material despite the AI-revenue mix; cohort gets mispriced on BTC-volatile days use the dislocation as an entry window, not an exit signal.
Pair-trade vs IREN APLD is the higher-beta, less-proven expression of the same neocloud / BTC→AI thesis.
CoreWeave counterparty concentration partly mitigated by the investment-grade SPV conversion and new hyperscaler tenants at Forge 2/3.
Earnings: Q4/FY26 print (fiscal year ends May) expected ~late-July to mid-August 2026; no confirmed date yet, re-check weekly. No blackout currently.
2026-06-09 financing shift: $1.59B of 7.00% senior secured notes due 2031 replaced near-term equity-ATM dilution risk with ~$110M+/yr interest and leverage. A subsequent equity-ATM raise would be a negative balance-sheet signal.
Backlog stepped to the high-$30Bs after Delta Forge 2 (~$5.2B, 210MW, 15-yr IG hyperscaler, 2026-06-08) on top of Polaris Forge 3 ($7.5B/300MW) and Forge 1 (CoreWeave, ~$11B/400MW, IG SPV).
Higher-beta, less-proven pair vs IREN on the same BTC→AI thesis; cohort registry flipped to MATURING 2026-06-07 even as company-specific flow accelerated.
Next refresh: pin exact levels 2026-06-05 low ($39.92), the rising 20-EMA and 50-EMA, and the post-06-09 surge high for machine-checkable invalidation.
Residual BTC beta remains material; use BTC-volatile dips toward the 20-EMA as entry windows, not exit panic.
Do not chase the 06-08/06-09 lease-news spike; cleaner risk/reward is a 20-EMA retest.
Next earnings: Q4/FY2026, est. 2026-08-10 to 08-20 (unconfirmed); fiscal year ends May. No earnings blackout inside the next 30 days as of 2026-07-11, but re-check for a confirmed date weekly.
Setup reframe 2026-07-11: price fully retraced the June spike ($50.73 high → $31.15 close 07-10). The June 'stretched above MAs' risk is gone, replaced by a good-news-sold divergence. Do NOT treat as a momentum long until price reclaims the mid-$30s→low-$40s MA zone on volume.
Real public campus names are Polaris Forge 1/2/3 use these, not internal aliases. PF1 = 400 MW, CoreWeave-anchored, ~$11B; PF2 = ~$5B investment-grade hyperscaler lease; PF3 = >1 GW milestone with MDU power deal 2026-06-22.
Backlog: ~$31B baseline contracted lease revenue, up to ~$73B with renewals across ~1.2 GW of IT load.
Build is now debt-funded ($1.59B 7.00% senior secured notes due 2031, priced 2026-06-09). Watch for any follow-on equity ATM raise as a balance-sheet stress signal.
Residual BTC beta still material (2026-06-24 crypto selloff dragged it lower); cohort trades it as a crypto proxy on risk-off days use flushes as entry windows only once structure firms, not exit panics.
Pair/cohort: IREN is the more-proven BTC→AI pivot and the natural pair. CoreWeave is both peer and PF1 anchor counterparty = concentration risk cuts both ways.
Earnings blackout ACTIVE: fiscal Q4/FY2026 results 2026-07-27 after close, call 5:00pm ET. Fiscal year ends May 31. No fresh entries inside the print window.
Archetype: Legacy Pivot confirmed: revenue inflection printed 2026-04-08, TTM revenue $319.3M +66.1%.
Structural downtrend since 2026-05-28 closing high $49.65. Price below 20/50/100-day averages as of 2026-07-17 at $25.79.
Key levels to re-pin each refresh: $25.56 April base floor, $31 broken shelf (lost 2026-07-10), $20 next visible support, 52w range $9.79-$50.73.
Hard rule: NEVER average down. The name is in a confirmed downtrend; size does not increase on weakness. Re-engagement only on a weekly close back above $31 with a post-print catalyst.
Beta 5.68 - position sizing must account for extreme volatility even at LOW conviction.
Debt stack ~$2.7B including $1.59B 7.00% senior secured notes due 2031 (priced 2026-06-09). One Polaris Forge 1 tranche still unplaced - watch for equity/ATM substitution as the primary bear trigger.
Nvidia sold all 7,716,050 shares (~$177M) per 13F covering 2025-12-31. Strategic validation anchor gone, not replaced.
Analyst gap is extreme: 10 Strong Buy, avg target ~$76.70 vs $25.79 quote. Zero Hold/Sell ratings means the downgrade path is entirely unpriced.
Pair-trade reference vs IREN: APLD is the higher-beta, less-proven expression. The APLD/IREN spread reads whether the market is discriminating on execution or selling the cohort.
BTC beta still material (confirmed 2026-06-24 crypto-linked selloff) despite AI revenue mix.

Related · shared themes

CRWV

CoreWeave, Inc.

Neocloud de-rating is confirming, not contested: Meta's reported ~$10B Anthropic compute deal (7/18) turns the disintermediation threat into a signed transaction, shares are -30% in a month, and desks are cutting (Mizuho $100, Macquarie $130). Broken structure into the 8/11 print stand aside until it bases.

LOW

IREN

IREN Limited

Neocloud re-rate has fully round-tripped: $33.62 close 7/17, -42% in 30 days, late-May Microsoft breakout base surrendered. Contracts intact ($9.7B MSFT, prepaid tranches) but Q3 AI revenue was only $17.3M vs a $3.4B ARR ambition, and Nvidia/Meta moving into direct compute sales plus the $800M CEO grant fight have flipped the theme to a live de-rating.

LOW

WULF

TeraWulf Inc.

Bitcoin-miner-to-HPC re-rate confirmed: Anthropic's 20-year ~$19B AI-infrastructure lease (2026-07-06) is the named-tenant catalyst the tape was missing; theme ACCELERATING with a wall of raises (MS $72, 2026-07-08), but a ~$3.5B financing overhang and a post-spike extended candle make a fresh chase the risk.

MEDIUM

HUT

Hut 8 Corp.

Bitcoin-miner-to-AI-datacenter-landlord re-rate is now financed fact ($4.25B IG notes 6/4, ~$16.8B contracted backlog), but the momentum leg rolled from ~$127 to ~$105 as peers WULF/IREN captured the marquee Anthropic tenant deals (7/6) and crypto weakened. MATURING, not accelerating needs its own named-tenant catalyst or a base reclaim before the next leg.

LOW