Applied Digital is a BTC miner repricing into a hyperscaler-grade datacenter operator. Two 8-Ks, seven days apart: the April 16 filing granted fully-vested executive MIP Units ahead of the ChronoScale/EKSO close; the April 23 filing disclosed a 15-year lease with a "U.S.-based high investment-grade hyperscaler" at Delta Forge 1. Stock has run +690% in twelve months to $35.89, and consensus is catching up. What shipped alongside the good news is the point of this memo.

What the filing says

The April 23 8-K (Item 7.01) disclosed a 300 MW, $7.5B, 15-year lease at Delta Forge 1 in Alexandria, Louisiana. This is APLD's third distinct hyperscaler tenant — CoreWeave at Polaris Forge 1, an unnamed "investment-grade hyperscaler" at Polaris Forge 2, and now an unnamed "high investment-grade" hyperscaler at DF1. Total contracted revenue crossed $23B, with >50% IG-backed. Target NOI margin 85% ±3%; CapEx $11–13M/MW. APLD simultaneously announced two $300M senior secured facilities — a bridge for Building 3 construction and a platform revolver — "expected to close promptly."

The April 16 8-K was narrow but informative. APLD granted fully-vested Management Incentive Plan Units to four executive officers, representing 5.25% of the EKSO shares to be issued to APLD's HoldCo. Lawyers don't issue vested equity in a subsidiary unless a deal is in final documentation. ChronoScale/EKSO closes imminently.

The less-flagged item: Exhibit 99.2 of the April 23 filing — an investor presentation — quietly revised Polaris Forge 1 Building 3 from "CY2026" (as the Q3 FY26 10-Q had guided on April 8) to "Calendar 1H 2027." A ≈6-month slip on the principal near-term revenue catalyst. Not headlined.

What the market thinks

At $35.89, APLD has largely priced the reclassification story — miner-to-operator multiple rerate from 5–8x EBITDA to 15–20x. A rough fundamental frame: $23B contracted revenue over ≈15 years at ≈86% blended NOI margin implies ≈$1.3B NOI at full ramp. Fifteen to twenty times that, discounted back five years at 15%, lands $30–50/share of intrinsic today. Current price sits inside that band.

Options structure tracks. Max pain $27, roughly 25% below spot. OTM put skew +41.7% versus flat call skew. IV inverted (near ≈60, mid ≈107) — the market expects vol expansion. 29.4% short interest. Beta 7.27. Dealer gamma pulls toward max pain absent a catalyst.

Scenario-weighted across five outcomes (bull execution 30%, base muddle 35%, mild setback 20%, material setback 10%, Base Electron fires 5%), expected 12-month return is +8% at $35.89 — Sharpe ≈0.22, below sizing threshold. At $30 entry, expected return is +30%, Sharpe ≈0.75. The entry level — not the directional call — is where the edge sits.

Why the gap exists

Three specific items are not in consensus:

1. Building 3 6-month slip. Sell-side models anchored on the April 8 10-Q. The April 23 investor deck revised the timeline; that revision is in an exhibit, not the press release. FY27 H1 HPC revenue estimates look ≈$50–75M too high.

2. DF1 tenant identity. Cross-ticker utility mapping eliminates Meta (AA- rating, wrong utility — Richland/Entergy) and Amazon (already committed to Caddo via SWEPCO/STACK). DF1 sits in Cleco service territory. Most likely: Microsoft (AAA, no announced Louisiana site) or Alphabet (AA+, vertically integrating via $4.8B Intersect + $9.9B PPA). Identity surfaces via Cleco LPSC rate filings for a 300 MW interconnect — typically 12–18 months before energization. A MSFT or GOOGL confirmation would compress 75–125 bps on the next note issuance.

3. Base Electron $2.4B guarantee. Related-party IPP formed by APLD executives. Off-balance-sheet contingent, $2.4B unconditional/irrevocable against APLD's $1.73B cash. Neither April 8-K touched it. The Building 3 slip may be its downstream symptom.

Risks, ranked

1. Base Electron trigger. Binary; structural impairment (-50% to -70%).
2. Building 3 second slip. Past 1H 2027 → sell-side FY28 cuts.
3. Beta 7.27. A 10% SPX drawdown translates roughly 1:7.
4. Momentum unwind. +690% 1Y, RSI 76. Crash risk is real, not decorative.
5. Hyperscaler capex pause. $700B at peak; +50% YoY is a steep curve to sustain.

Catalysts

• ~May 31, 2026: ChronoScale/EKSO close (78% probability)
• Late July / early August 2026: Q4 FY26 earnings, first GAAP-clean quarter
• ~August 31, 2026: $600M bank facilities close (85%)
• 2H 2026: PF2 Building 1 (150 MW) initial operations
• By December 31, 2026: DF1 tenant identity publicly confirmed (70%)
• By June 30, 2027: Building 3 operational on revised guidance (65%)
• Anytime: Base Electron 8-K (unpredictable, binary)

What would change our mind

• Base Electron 8-K disclosing independent financing or IPO (bullish — exits tail risk), or a guarantee trigger (bearish — thesis break)
• Building 3 slipping a second time past 1H 2027
• ChronoScale/EKSO close delayed past June 30, 2026
• DF1 tenant identity reveals as BBB (Oracle) or weaker IG (Meta AA-)
• Any insider Form 4 open-market sale >$5M at current levels

Evidence