BlitzLab Scoring · Oct 2026 · Model v1.0 (frozen)

Cipher Digital — Blitz Index 32 (D)

Sold before built — the presale blitz of the AI era.

32Grade D

Confidence: Medium · Category: ai-infra

Score breakdown

Section	Score	Detail
A · Speed	10 / 25	Recognized revenue declining (Q2 −43% YoY); TTM $191M vs $11.4B contracted
B · Moat	24 / 60	700MW HPC contracted; Black Pearl first delivery Aug 2026; 5.3GW pipeline
C · Creative destruction	9 / 15	Landlord layer — hyperscalers sign 15-year leases with ex-miners
D · BlitzFailing	−11	D1 full trigger (−10): borrowings $5.54B = ~29x TTM; negative interest coverage

Radar axes (0–100): Moat 40 · Speed 40 · Creative 60.

Watch item

Gates: G1 pass (conditional) · G2 pass (estimated). 33 below threshold — first-ever score. Second D grade after DBX (34/D). Renamed Cipher Mining → Cipher Digital (Feb 2026).

Strengths

Contracted revenue ~$11.4B (~60x TTM) — 10–20yr take-or-pay (SEC for $9B+ Barber Lake)
Black Pearl delivered early (Aug 2026), rent recognition started — first execution proof
5.3GW pipeline across 11 sites — power is the moat in this market
Google backstop on Fluidstack lease ($544.5M) — credit enhancement (SEC)

Concerns

Borrowings $5.54B = ~29x TTM, negative interest coverage — worst leverage of the four
Recognized revenue declining; gross profit and adj. EBITDA negative — no unit economics yet
Barber Lake cost overruns ($359.3M) borne by CIFR — direct construction exposure
HPC revenue in no filed results yet — Q3 2026 is the first test

Research notes

1. Current Position — Act 1 of the Pre-Sale Blitz (Contracts Are Done; Delivery Is the Question)

CIFR is **Act 1 of the "pre-sale blitz."** The model of selling apartments before they're built — the selling (contracts) is done, and move-in (delivery) is the question. One number summarizes this company's position: **TTM revenue of $191M (SEC confirmed, all bitcoin mining, shrinking) vs contracted revenue of ~$11.4B (press-reported) — a 60x gap.** The gap is everything. Demand is confirmed in SEC documents (Barber Lake $9B+, 20 years, 2026-09-25 8-K). But the revenue is not yet on the P&L — **HPC revenue has never appeared in any filed results.** The Q3 earnings are the first proving ground for HPC revenue. v1.0 looks only at recognized numbers — A1 2 points, B4 2 points. **GATE verdict:** G1 PASS (conditional · estimate) — a "contracted curve," not a "recognized curve." G2 PASS (estimate) — ~6x the bar of $1.91B on contracted revenue of ~$11.4B alone. The weakest G1 pass in coverage.

2. Blitzscaling Lens

**Rule 8. "Raise Too Much Money" — but speed was bought with debt.** CIFR is Rule 8's most expensive execution. Black Pearl $2B HY bonds (2026-Q1, press-reported) + Stingray $810M 6% bonds (2026-06, press-reported) + Morgan Stanley $200M revolver (SEC confirmed) — **~$6B raised over 13 months (press-reported)**, at 6%+ high-yield pricing. NBIS's opposite — CIFR is the extreme form of "building with borrowed money." Financing speed is fast but expensive, and dilution continues (shares outstanding 405.0M → 414.3M, SEC).

3. Competitive Landscape — Two Species of Neocloud

**The designated competitors** are the cohort's two species. The GPU-cloud type (IREN·CRWV·NBIS — selling GPUs themselves) and the hosting/campus developer type (CIFR·APLD·WULF·CORZ — building and leasing data centers). CIFR is plotted as "the fastest AI DC conversion" in the second species (provided candidate memo).

4. Inflection Points

**Inflection Point 1. Q3 earnings — the first proving ground for HPC revenue.** What to watch: whether **HPC segment revenue registers in filed results for the first time** at the Q3 2026 earnings. Black Pearl rent recognition started 2026-08 (SEC confirmed) — Q3 is the first quarter where HPC revenue is measured. The YoY direction of recognized revenue (Q2 −43% YoY) and whether adjusted EBITDA (−$30M per quarter) improves are the observation points. As long as A1 2 points · B4 2 points are tied to "recognized numbers," this inflection point is the only door that moves the score. **Earnings-date conflict:** TipRanks 11/11 pre-market "Confirmed" (press-reported) vs MarketBeat 11/2 estimate (press-reported) vs marketchameleon's 10/30–11/4 window (press-reported), no company announcement — no date asserted. Link: before v1.1 candidate 1 ("A1 reflecting the contract/recognition gap ratio") is formalized, the $11.4B of contracts is not directly reflected in the score. v1.0's structural blind spot and the reason for 32/D. The Q3 earnings shine the first light into that blind spot.

5. One-Line Risk

In a structure with $5.54B of borrowings (29x TTM) and negative interest coverage, if the delivery speed of the $11.4B contracts cannot outrun the accumulation speed of interest — as WeWork's lesson teaches, the real crisis comes when execution, not demand, breaks. "Delivery is everything."

1. Current Position — Act 1 of the Pre-Sale Blitz (Contracts Are Done; Delivery Is the Question)

But the weight of 32 is in D. **Borrowings of $5.54B (LT $5.45B + ST $0.09B, 6/30 SEC confirmed) = ~29x TTM revenue — the worst in the neocloud cohort** (NBIS ~10x, CRWV ~4.6x). Interest of $66.7M in Q2 (annualized ~$267M, SEC) vs adjusted EBITDA of −$30M per quarter — negative interest coverage. Cash of $0.83B, accumulated deficit of −$1.39B (SEC confirmed). The D1 full trigger of −10 is the single largest variable of the CIFR scoring. Total deductions of 11 points sit 4 points below the FLAG bar (15 points) — a "watch" band, not a "warning" band.

**Two-sided assessment (per the score file):** - **On the upside:** Contracted revenue of ~$11.4B (~60x TTM) — 10-to-20-year take-or-pay, AWS·Fluidstack·"leading AI lab"·third hyperscaler (some counterparty names press-reported) · Barber Lake $9B+ (20 years, SEC confirmed) + Google backstop (SEC confirmed) · Black Pearl's early August delivery and rent commencement (SEC confirmed) · 5.3 GW pipeline (11 sites, SEC 10-Q confirmed) · name change (Cipher Mining → Cipher Digital, 2026-02, press-reported) — the declarative confirmation of the strategic transition. - **On the downside:** Recognized revenue shrinking (Q2 −43% YoY), negative gross profit, adjusted EBITDA of −$30M per quarter · borrowings of $5.54B (29x TTM) + negative interest coverage · Barber Lake's $359.3M of cost overruns borne by CIFR (only 50% recovered beyond that) — direct exposure to construction overruns · effectively all contracts concentrated in 3–4 tenants · Bitfury affiliates holding 19.4% (13D/A 2026-10-07, confirmed).

**Valuation note (not reflected in the score):** market cap of ~$5.5–7B (press-reported, varies by time point), P/S ~29–37x (on TTM revenue of $191M), EV of ~$11–12B (estimate) including $5.54B of borrowings. The business score does not look at price (the ARK tests' lesson).

The directional assessment is "pre-sale done, move-in pending" — the score has exactly one direction: the delivery speed of the $11.4B contracts.

2. Blitzscaling Lens

This rule's sharp point is the "quality" of financing. NBIS's money is customer prepayments; IREN's money is contract-backed project financing. CIFR's money is high-yield bonds from the debt market — Q2 interest of $66.7M accumulates to ~$267M annualized. At 29x leverage, one quarter of delivery delay carries a different weight than one quarter's miss at a software company: **debt interest accrues daily regardless of delivery.** That is the background of v1.1 candidate 2 ("D1 sub-segmentation of project-finance-type debt"). Campus-level nonrecourse financing is a structural buffer, but v1.0 assigns −10 on the aggregate basis.

**Rule 5. "Let It Burn" — the big fire is delivery (construction); everything else is a small fire.** Hoffman's fire priority is "extinguish only the fires that can kill the company; let the small ones burn." For CIFR the true big fire is **whether the $11.4B promise can be built** — construction, delivery, power grid interconnection. The condition that CIFR bears Barber Lake's $359.3M of cost overruns (SEC 8-K) shows exactly where this big fire sits: if the building goes sideways, the cost is entirely the company's. Black Pearl's two-month-early delivery was the first case of putting out this big fire.

Here the "small fires" are large. Recognized revenue −43% YoY, negative gross profit, adjusted EBITDA −$30M — everything looks red, but in Act 1 of the pre-sale blitz these are "small fires." As the score file judges: "efficiency is a later problem; the current negative adjusted EBITDA is the textbook loss phase of blitz Act 1." But there is a trap: **in a debt-built structure the small fires do not go out on their own** — the fuel of interest keeps flowing. The essence of this model is "demand risk lowered, converted into execution risk."

**Revisiting the applicability conditions — blitzscaling without network effects.** Among the 9 rules' applicability conditions, "a business model where the rewards of speed are large (network effects, virality, data advantage)" — CIFR does not qualify (B1 5 points, "no true network effects"). Tenants buy power and sites, not CIFR. The winner-take-all reward lies not in software's network effects but in preempting **"AI-era distribution" — power and sites** — the grammar of the "land-grab" family with APLD and WULF.

Hence the score file's verdict of "the AI era's real-estate developer." A different species from the three-way neocloud race (CRWV·NBIS·IREN — those who sell GPU clouds). B1 5 points · C1 3 points are this difference's score — there is contractual lock-in of 10-to-20-year take-or-pay long-term leases + the Google backstop ($544.5M, SEC confirmed), but no compute-stack switching costs or margin structure (the Google backstop is "credit enhancement, not a moat").

3. Competitive Landscape — Two Species of Neocloud

**The intra-species comparison axis** is this research's core. APLD has contracted revenue of ~$36B (1.41GW+, 15-year take-or-pay, press-reported) — more than 3x CIFR's ($11.4B). WULF leads on marquee-customer quality with the Anthropic 20-year lease ($19B of contracted revenue, press-reported). CIFR's distinction is the combination of a single massive confirmed contract — Barber Lake $9B+ (20 years, SEC confirmed) — and the execution proof of Black Pearl's early delivery. But it is also clear that **CIFR is the weakest link within the species**: the worst leverage of the four neoclouds (29x), the smallest recognized-revenue base ($191M TTM), and only 700MW of actual HPC contracts out of the 5.3 GW pipeline.

**The IREN (56/C) comparison** is the contrast of "same miner origin, different pivot." IREN went to the vertically integrated AI cloud (the species that sells compute); CIFR went to the campus developer (the species that leases buildings). IREN's Microsoft $9.7B contract (confirmed) competes for the same customer pool as CIFR's backlog — the hyperscalers' demand wave reaches both species. But hyperscalers are also "customers and structural landlord candidates" for CIFR: AWS is a tenant (press-reported) while building data centers itself. C1 3 points is this dual structure's score.

**CORZ's warning** is this entire species' memory. CORZ (contracted capacity ~1.1GW, base contracted revenue $24B+) carries the 2022-12 Chapter 11 filing → 2024-01 revival history (candidate memo). A precedent showing where high capex · losses · execution delays lead — a coordinate to read alongside CIFR's D1 −10. **When the demand wave breaks, debt-built buildings collapse first** — that is CORZ's lesson.

**The structural competitor** is ultimately construction execution itself. The bottleneck is not GPUs but grid interconnection, power delivery, and construction schedules — the B2 discount factor ("the contracts are sold, the buildings are under construction") is this species' shared grammar. If the phased Barber Lake delivery (Q4 2026–Q1 2027) is confirmed by measurement, B2's discount lifts; if overruns beyond the $359.3M cap occur, D1 −10's warning converts into real liquidity pressure.

4. Inflection Points

**Inflection Point 2. Barber Lake delivery — the $359.3M test.** What to watch: schedule adherence of the phased Barber Lake delivery (Q4 2026–Q1 2027, SEC confirmed). The $359.3M of cost overruns borne by CIFR (only 50% recovered beyond that) is **direct exposure to construction-overrun risk**. Black Pearl was two months early, but Barber Lake is the core campus of the $9B+ 20-year contract — a different scale. Whether overruns stop at the $359.3M line is the observation point. Link: the inflection point of whether the B2 8-point discount factor lifts or deepens. Barber Lake is the next wave of the 700MW of actual HPC contracts (3 sites) out of the 5.3 GW pipeline (11 sites). One quarter of delivery delay meeting $66.7M of quarterly interest (SEC) shakes the whole leverage structure — in the heavy blitz, delay is cost.

**Inflection Point 3. The leverage firewall — the direction of interest coverage.** What to watch: the quarterly 10-Q ratio of interest expense (Q2 $66.7M) to adjusted EBITDA (−$30M per quarter) — **the interest-coverage direction is this company's lifeline**. Cash burn against cash of $0.83B, and whether additional financing terms deteriorate — observed together. If any one of delivery delays, rate hikes, or tenant disruptions fires, liquidity pressure is direct. Current deductions of 11 points sit 4 points below the FLAG bar (15 points) — a "watch" band. Link: campus-level nonrecourse project financing as a structural buffer (press-reported) is a plus factor for the firewall, but v1.0's D1 assigns −10 on the aggregate number. Until v1.1 candidate 2's sub-segmentation arrives, the number $5.54B itself is the score's ceiling.

**Observation checklist (quarterly):** ① first measured HPC revenue at Q3 earnings + recognized-revenue YoY direction (earnings-date conflict — TipRanks 11/11 "Confirmed" / MarketBeat 11/2 estimate / marketchameleon 10/30–11/4 window; no company announcement) ② phased Barber Lake delivery (Q4 2026–Q1 2027) schedule + adherence to the $359.3M overrun cap ③ direction of the interest-expense/adjusted-EBITDA ratio (Q2 $66.7M vs −$30M) ④ quarterly burn against cash $0.83B ⑤ expansion/maintenance of contracted revenue ~$11.4B (press-reported). ①·②·⑤ are "execution" metrics; ③·④ are "firewall" metrics. If execution metrics trend up and the firewall holds, 32/D rises in the directional direction. If delivery delays meet overrun-cap breaches, D1's (−10) warning converts into real leverage pressure — in a debt-built structure, delayed promises are increased costs.

5. One-Line Risk

*BlitzLab v1.0 (frozen). Unconfirmed figures must be labeled estimate / press-reported.* *For education and research purposes only. Not a recommendation to buy, sell, or hold any security.*

### Verification Log - Verification review waived.

*Translated from the Korean edition (2026-10-15). For education and research purposes only. Not a recommendation to buy, sell, or hold any security.*

Next earnings

Nov 2, 2026 · Expected

Coverage

As of October 2026 · Model v1.0 (frozen) · Confidence: Medium (listed, SEC filings) · Not investment advice — a scholar's assessment.

BlitzLab scores businesses against a scholar’s standard — not investment advice. No price targets, no buy/sell recommendations. Figures are from SEC filings (XBRL) as labeled in the source scorecard.