Hut 8

Phase A — Understand the business

Lens 1 · Company Overview

Hut 8 is no longer a bitcoin miner. As of the FY2025 10-K it self-describes as "an energy infrastructure platform that integrates power, digital infrastructure, and compute at scale" with a "power-first" development model. The reality the filings reveal is sharper than that marketing line: HUT has become a developer-landlord of hyperscale AI data centers with a large bitcoin treasury bolted to its balance sheet.

The platform is three stacked layers:

• Power — owns/develops/manages powered land, interconnects, substations, switchyards. 1,020 MW under management across 15 sites in the US and Canada at YE2025.
• Digital Infrastructure — ASIC hosting/colocation + CPU cloud (Hut 8 Canada, 200+ customers, 5 Canadian data centers, ~3 MW, 36,000 sq ft) + the new AI data centers (River Bend, Louisiana; Beacon Point, Texas).
• Compute — three sub-lines: ASIC Compute (bitcoin mining, now run almost entirely through majority-owned American Bitcoin Corp / ABTC), Traditional Cloud (Hut 8 Canada), and AI Cloud (Highrise AI, Inc. — 1,000 NVIDIA H100s + 96 H200s at a Chicago colo).

Four reportable segments: Power, Digital Infrastructure, Compute, Other.

How it actually makes money today vs. tomorrow — this is the whole story:

• Today's revenue ($71.0M in Q1-2026) is ~93% Compute, i.e. bitcoin mining rewards earned by ABTC. The P&L is still a bitcoin-mining P&L.
• Tomorrow's revenue is two signed, triple-net, long-dated leases to hyperscale-grade tenants that have not yet started paying:
• River Bend (Louisiana) — 15-yr triple-net lease to Fluidstack for 245 MW, base ~$7.0B / up to $17.7B with renewals, with the lease payments backstopped by Google for the 15-yr base term; part of a partnership with Anthropic to jointly develop up to 2,295 MW total. First data hall Q2 2027.
• Beacon Point (Texas) — 15-yr triple-net lease to a multi-trillion-dollar-market-cap, high-investment-grade technology company, base ~$9.8B / up to $25.1B with renewals, ~$655M average annual NOI, 3% annual escalators, initial 352 MW IT load, delivery Q3 2027. June-2026 bond docs say the tenant is rated AA- or higher.

Contract structure / key terms: triple-net (tenant bears opex/taxes/insurance), 15-year base with multiple 5-year renewal options, fixed escalators — i.e. REIT-like contracted cash flows, not commodity exposure. The bitcoin-mining revenue, by contrast, is fully variable on BTC price and network difficulty.

Main counterparties named in filings: tenants Fluidstack + the unnamed AA- tech tenant; Anthropic (AI partner) and Google (River Bend backstop); Bitmain (miner supplier, paid in pledged BTC); power providers Entergy and Vertiv; J.P. Morgan / Goldman Sachs (River Bend loan underwriters); Jacobs (GC); custodians Coinbase, NYDIG, Anchorage, BitGo.

Lens 2 · Supply Chain

supply-chain.md is missing from the commercial layer, so this is built from the filings. The chain runs power → infrastructure → compute → end-customer, and HUT's deliberate strategy is to own the scarce upstream (power/land) and rent out the midstream (the building):

Upstream — power & land (the real moat input):

• Electrons: ERCOT grid + wind (Texas — Vega, Salt Creek, Beacon Point, River Bend pipeline), AESO + CCGT (Alberta — Medicine Hat, Drumheller), Entergy (River Bend power).
• Powered land: owns 627 acres (Louisiana / River Bend), 524 acres (Nueces County, TX / Beacon Point), 20 acres (Medicine Hat), plus Salt Creek substation land and Batavia, IL. Chokepoint = interconnect queue + power availability; HUT's pipeline metric ladder (diligence → exclusivity → development → construction → management) is explicitly a power-origination funnel.

Midstream — data-center build:

• Vertiv (power/cooling equipment), Jacobs (general contractor), Bitmain (ASIC miners). Bitmain is a notable single-source ASIC dependency, paid in pledged bitcoin rather than cash — e.g. 16,299 S21 miners (~14.02 EH/s) for ~$314M settled via 2,234 BTC pledged. Tariff risk on imported equipment is explicitly flagged.

Downstream — who pays:

• AI/HPC tenants (Fluidstack→Anthropic; the AA- Beacon Point tenant) — the future cash flow.
• Bitcoin network / mining pools — today's cash flow, via ABTC.
• Hut 8 Canada cloud customers (200+) and Highrise AI colo provider.

Single-source / concentration flags: Bitmain (miners), one Managed-Services site (280 MW at the King Mountain JV, 50%-owned), and — most importantly — two tenants will eventually be ~all of the contracted backlog. Names along the chain are present, so this lens is grounded, not generic.

Lens 3 · Competitive Advantages (moats)

The moat is not mining efficiency (a short-seller once called HUT "the most inefficient bitcoin miner" — see Lens 9/10). The durable moat is power origination + a credibility flywheel into investment-grade capital markets.

1. Powered land + interconnect rights at scale. HUT discloses 5,185 MW under diligence, 1,755 MW under exclusivity, 1,230 MW under development, 330 MW under construction at YE2025. In a world where the binding constraint on AI is megawatts, not GPUs, controlling a multi-GW interconnect pipeline is the scarce asset. This is a real, hard-to-replicate moat.
2. Investment-grade project-finance access — uniquely demonstrated. River Bend's April-2026 $3.25B bond was "the first investment-grade project bond ever issued for a construction-stage data center project" (BBB−, S&P/Fitch). Beacon Point's June-2026 $4.25B bond priced at Baa2/Moody's, 20bp inside River Bend. Cumulative IG project-level financing $7.5B. The ability to fund builds with non-recourse, asset-level IG debt — rather than dilutive equity — is a structural cost-of-capital edge most miner-peers cannot match.
3. Blue-chip counterparty validation. Google backstop + Anthropic partnership + an AA- tenant + JPM/GS underwriting is a reputational moat: it lets HUT win the next tenant and the next bond more cheaply (a flywheel).
4. Bargaining power: weak vs. hyperscaler tenants (they have many landlord options and dictate credit terms), but strong vs. capital markets now that the IG-bond template is proven, and strong vs. power counterparties where it already holds the land/queue position.

Net: a genuine moat, but it is a real-asset / cost-of-capital moat, not a technology moat — and it is young (the leases haven't started paying).

Lens 4 · Segments

segments.csv is empty; figures below are pulled from the Q1-2026 income statement. Revenue is disaggregated into Power / Digital Infrastructure / Compute (intercompany ABTC fees eliminated in consolidation):

[YoY =: e.g. Compute 65,974/16,118−1 = +309%; total 71,017/21,815−1 = +226%]

Read of the trend:

• Compute is the entire growth story and ~93% of revenue — and it is essentially ABTC bitcoin-mining rewards. So "revenue +226%" is mostly more bitcoin mined at a higher BTC price + a bigger fleet, not the data-center pivot showing up yet.
• Power fell because the Far North JV (310 MW Ontario gas plants) was divested in Feb 2026 — a deliberate exit of merchant power generation.
• Digital Infrastructure is tiny and flat — the legacy hosting/cloud book. The River Bend / Beacon Point leases will land here (and in Power) starting 2027; today this segment understates the future by an order of magnitude.
• Geography: US-weighted (Texas/New York) plus Alberta (Canada); the Canadian cloud book is small. A "significant portion" of BTC is held by the Canadian subsidiary (FX exposure).

The segment table is the trap with this name: the reported mix says "bitcoin miner"; the contracted mix (Lens 1) says "AI landlord." Both are true; the gap is timing (2027).

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print — Q1 2026, reported 2026-05-06)

Headline:

• Revenue $71.0M (+226% YoY).
• Operating loss $(370.4)M vs $(147.7)M — but this is dominated by a $295.7M non-cash "Losses on digital assets" mark (BTC price fell intra-quarter) plus $81.7M G&A (which includes $50.9M stock-based comp) and $38.4M D&A.
• Net loss $(253.1)M; net loss attributable to Hut 8 $(219.8)M; EPS $(1.98) basic & diluted (111.1M wtd shares).
• Below the line: a $33.6M gain on the Far North JV sale, $40.8M gain on derivatives, and a $48.9M income-tax benefit.

Gross economics are actually fine — total cost of revenue was only $25.5M on $71.0M revenue (~64% gross margin ex-D&A). The optical loss is the digital-asset mark (the company carries BTC at fair value, so a down-quarter in BTC slams the GAAP P&L) plus a SBC spike. The loss is a mark-to-market loss, not a cash-burn loss.

Balance sheet (3/31/26):

• Total assets $2.61B, total liabilities $919.6M, total equity $1.69B (of which $310.3M is non-controlling interest = ABTC minority).
• Cash $160.0M (up from $44.9M at YE2025, on ATM issuance + ABTC raises).
• Bitcoin on B/S ≈ $1.124B: held-in-custody $435.7M + pledged-for-miners $210.8M + pledged-as-collateral $477.5M.
• PP&E $812.4M, Goodwill $209.2M, Derivative assets $161.7M (non-current).
• Debt: loans/notes/other financial liabilities $404.8M ($200.0M current + $204.8M non-current) + miner purchase liability $360.9M + operating leases. (Total debt was ~$411.1M at YE2025, before the $3.25B + $4.25B project bonds drawn in Apr/Jun 2026.)

Guidance/tone: no formal numeric guidance (guidance.csv empty; HUT doesn't guide EPS). MD&A tone is confidently expansionary — management says liquidity from "cash flows, Bitcoin on the balance sheet, and other financing sources will be sufficient" and that the builds will be funded by project-level financing, not corporate equity. No going-concern flag.

Market reaction / context: Q3-2025 (the prior print) was profitable — $83.5M revenue, $50.6M net income, $109.0M adjusted EBITDA. So the quarter-to-quarter GAAP swing is almost entirely the BTC mark flipping sign — the textbook tell that this equity is a levered bitcoin call with a data-center option stapled on.

Unusual vs. its own history: the $50.9M SBC (vs $3.8M a year prior) is a step-change — driven by performance-stock-units tied to commercializing capacity / earnings / operational milestones. Watch dilution.

Lens 6 · Earnings Calls (sentiment trend)

transcripts/ is empty; this is ``-grounded and necessarily lighter. The arc across the last ~4 quarters is a clear, deliberate narrative migration from "bitcoin miner" → "energy/AI-infrastructure platform":

• Early/mid-2025: language centered on fleet upgrades (S21+ miners), ASIC efficiency, and launching ABTC as the bitcoin-accumulation vehicle.
• Q4-2025 / 2025 highlights: the pivot crystallizes — Anthropic/Fluidstack/River Bend announced Dec 2025; Far North sale to TransAlta; pipeline expanded to thousands of MW.
• Q1-2026 + after: the dominant message is "power-first," triple-net leases, NOI, investment-grade project finance, non-recourse, disciplined corporate leverage — the vocabulary of a data-center REIT/developer, not a miner.

What they stopped saying: hash-rate bravado as the headline. What they started saying: "multi-trillion-dollar tenant," "NOI," "cost of capital at the asset level," "Bitcoin per Share" (the ABTC framing). Sentiment = confident/expansionary; the risk is that confidence is running ~5 quarters ahead of revenue.

Lens 7 · Comps

Multiples below are ``; where a clean multiple wasn't sourced I mark n/a rather than fabricate. The peer set is the bitcoin-miner-to-AI-infrastructure cohort.

Sector valuation frame: miners with secured HPC contracts trade ~12.3× NTM sales vs pure-play miners ~5.9×; the bull path is convergence toward data-center-REIT multiples (>20× EV/EBITDA) as contracted cash flows season. 5-yr avg ROE: n/a — not meaningful (these names have volatile, BTC-mark-driven, mostly-negative GAAP earnings histories; a 5-yr ROE would be noise). Dividend yield: 0% across the cohort (HUT has never paid a dividend ).

Honest read: HUT screens as mid-cap within the cohort (~$14B, between RIOT ~$6B and IREN ~$18B), and arguably the highest-quality contracted backlog (Google/AA- credit, IG bonds) — but I will not manufacture an EV/EBITDA multiple the sources didn't give. The defensible statement is relative: HUT's backlog quality > peers; its multiple is set by how much of that 2027 cash flow the market chooses to capitalize today.

Lens 8 · Stock-Price Catalysts (what actually moves HUT)

Pattern over the last ~2.5 years, `` + filing-dated:

• Bitcoin price — the dominant high-frequency driver. The GAAP P&L (and thus headlines) swing on the BTC mark; the equity behaves as a levered BTC beta.
• Short-seller report (Jan 18 2024) — −23% in a day on the JCapital Research report alleging pump-and-dump ties (Honig Group) via then-strategy-officer Michael Ho; CEO departed weeks later. This is the single largest idiosyncratic down-move and the origin of the current management regime.
• AI-lease announcements — River Bend/Anthropic (Dec 2025) and Beacon Point (early 2026) re-rated the stock from "miner" toward "AI infra"; The Block/CCN note HUT hitting fresh highs on the miner-to-AI boom.
• Investment-grade bond closings — River Bend $3.25B (Apr 30 2026) and Beacon Point $4.25B (Jun 9 2026) were de-risking catalysts (proof the builds are financeable non-dilutively).
• ABTC events — the Nasdaq listing (Sept 3 2025 via Gryphon merger) and ABTC's own BTC-per-share / capital raises move HUT via its ~80% stake.
• Sector/macro — semiconductor rallies and "data-center power crisis" narratives lift the whole cohort together; CCP/China BTC-crackdown headlines hit it.

Takeaway: the market reacts to (1) BTC price daily, (2) tenant/credit milestones structurally, (3) governance shocks violently. The 2027 lease commencement is the next big structural catalyst.

---

Phase C — Judge people & books

Lens 9 · Management

• CEO: Asher Genoot (age ~30). Co-founder & former COO/President of US Bitcoin Corp (USBTC), which merged with Hut 8 (the "Business Combination") to form today's company; became CEO Feb 6, 2024 — i.e. immediately after the JCapital short report and the prior CEO's exit. Track record: he is the architect of the power-first/AI-landlord pivot and the IG-bond strategy — which, if River Bend/Beacon Point deliver, is a genuinely impressive value-creation arc from a sub-$2 stock in early 2024 to ~$125 in mid-2026.
• Skin in the game: Genoot directly owns ~3.06% of HUT (worth ~C$225M); total insiders ~10.6%, institutions ~53%, retail ~36%. Decent founder alignment.
• Compensation: ~$10.3M/yr, ~95% equity/bonus, with 2024 LTI 100% in performance-stock-units (2-yr price measurement, 3-yr cliff). The huge $50.9M Q1-2026 SBC is the cash-flow-friendly but dilutive flip side of this design.
• Capital allocation: the defining recent moves are good — divesting merchant power (Far North → TransAlta, +$33.6M gain), refinancing 9.00% Coinbase debt into 7.00% FalconX, and (critically) funding the AI builds with non-recourse IG project bonds instead of equity. Against that: heavy ATM equity issuance ($120.1M in Q1-2026 alone ) and a bitcoin-treasury accumulation strategy that adds volatility.
• Board: Chair is E. Stanley O'Neal, former Merrill Lynch CEO — a heavyweight, capital-markets-credible (if historically polarizing) chair.
• Founder vs professional: founder-operator archetype (young, aggressive, equity-heavy, narrative-driven) — fitting for a company making a high-stakes platform pivot, but it raises the governance bar (see Lens 10/13).

Lens 10 · Forensic Red Flags

regulatory/regulatory-findings.md: 0 SEC Litigation Releases, 0 AAERs naming Hut 8 (EDGAR EFTS, 2021–2026). 10-K Item 3 Legal Proceedings: "not presently a party to any... legal or regulatory proceedings that... would... have a material adverse effect," with only a legacy USBTC rescission offer (July 2021, ~$5.87M notes) noted. Non-SEC web check: the material item is not a regulator action but the JCapital Research short report (Jan 2024) alleging pump-and-dump associations via Michael Ho / the Honig Group — HUT rebutted it as "inaccuracies and misrepresented data"; no enforcement action followed, but the then-CEO left. Treat it as a governance/reputation flag, not a confirmed accounting fraud.

Accounting risks to watch (every figure labeled):

1. Digital-asset fair-value volatility (income-statement noise). BTC carried at fair value → $295.7M "Losses on digital assets" in one quarter swamps a $71.0M-revenue P&L. GAAP earnings are nearly uninterpretable quarter-to-quarter; you must look through to cash and BTC count.
2. Stock-based comp flattering nothing — but diluting everything. $50.9M SBC in Q1-2026 (≈13× the prior-year quarter) with $144.4M unrecognized PSU expense to come. Non-GAAP adjusted EBITDA adds all of this back; the dilution is real.
3. Gains-below-the-line propping pre-tax results. $40.8M derivative gain + $33.6M JV-sale gain + $6.4M JV equity earnings = $80.8M of "other income" in Q1-2026. High-quality? The JV sale is one-time; the derivative gains are mark-driven (mirror of the BTC mark).
4. Goodwill $209.2M (from the USBTC/Gryphon combinations) — impairment risk if the equity de-rates.
5. Consolidation optics — ABTC. HUT consolidates 100% of ABTC's revenue and bitcoin but only owns ~80%; $310.3M of equity is non-controlling interest. Consolidated headline figures overstate the economics attributable to HUT shareholders. Intercompany MSA/CSA fees (the Power/Digital-Infra revenue from ABTC) are eliminated in consolidation — so the "infrastructure platform" revenue is largely invisible until third-party tenants (River Bend/Beacon Point) pay.
6. Bitcoin pledged everywhere. ~$688M of BTC is pledged (for miner purchases + as loan collateral); the FalconX loan has margin-call at 130% / liquidation at 105% — a BTC crash could force liquidations into weakness.

Verdict on the books: clean of regulators, but high-volatility and dilution-heavy accounting that rewards careful look-through. No going-concern, no fraud finding — but the GAAP statements are a poor guide; cash, BTC count, debt-maturity ladder, and contracted backlog are the real instruments.

Regulatory findings summary: No material SEC (LR/AAER) or 10-K Item 3 findings — verified via SEC EDGAR EFTS and the FY2025 10-K as of 2026-06-20. The only material adversarial event is the 2024 JCapital short report, which produced no enforcement action.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (FY2026 / FY2027 / FY2028 EPS)

No forecast.ts forecast logged (per --watchlist rules). GAAP EPS for this name is dominated by the non-cash BTC mark and SBC, so a point EPS estimate is low-information; I build the shape and label every input ``. Fiscal year = calendar year.

Anchor actuals: Q1-2026 revenue $71.0M; Q3-2025 (last clean profitable q) revenue $83.5M / NI $50.6M / adj EBITDA $109.0M; 15,679 BTC at YE2025; ~112.5M shares; two leases commencing Q2/Q3 2027.

• FY2026 (the "trough" year): GAAP EPS is un-forecastable as a single number — it will be whatever the YE BTC mark dictates. Operating/mining cash generation should stay positive (≈64% gross margin ex-D&A on rising mined revenue). Sell-side consensus is itself split — −$0.99 to +$1.38 depending on source — which is the honest answer: n/a as a precise figure; directionally near-breakeven-to-modest-loss on GAAP, positive on adjusted EBITDA, BTC-mark-dependent.
• FY2027 (inflection): River Bend (Q2) + Beacon Point (Q3) begin partial-year rent. Beacon Point alone targets ~$655M average annual NOI; River Bend base ~$7.0B/15yr ≈ ~$467M/yr. Combined run-rate NOI at full ramp ≈ >$1.1B/yr, but 2027 captures only a partial year + ongoing BTC-mark noise. EPS: still GAAP-volatile, but the first year the infrastructure earnings become visible.
• FY2028 (first full-year platform): first full year of both leases → contracted, escalating, triple-net NOI of ~$1.1B+/yr against project-level interest on ~$7.5B IG debt (~6.1–6.2% ≈ ~$460M/yr ). Net contracted spread ≈ $600M+/yr pre-corporate-cost — this is the number the bull case capitalizes. On ~112–125M shares (pre further dilution), that is a credible path to >$3–4/yr of platform earnings power, before any BTC-mark swings or ABTC NCI leakage.

The honest projection: GAAP EPS FY26–27 = n/a as a clean figure (BTC-mark-driven); the investable number is 2028 contracted NOI spread (~$600M+/yr ), which is what justifies a ~$14B cap today — i.e. the market is already paying for 2027–28 execution.

Lens 12 · Bull vs Bear

Bull case. HUT controls multi-GW of powered land/interconnect — the single scarcest input to the AI build-out — and has proven it can monetize it via two 15-year triple-net leases worth ~$16.8B base / ~$42.8B with renewals to Google-backstopped / AA-credit tenants, funded with $7.5B of investment-grade, non-recourse project bonds that don't dilute equity. From 2027 it converts from a volatile bitcoin miner into a contracted, escalating, REIT-like NOI machine while keeping ~15,679 BTC of upside optionality. If contracted-miner multiples (12.3× sales) drift toward data-center-REIT multiples (>20× EBITDA), the re-rate is large. The market is buying a hyperscaler landlord that happens to hold a bitcoin call.

Bear case (permanent-impairment risks).

1. Execution/timing gap. Zero of the $16.8B has been earned; first cash is Q2–Q3 2027. Construction delay, interconnection slippage, or a tenant dispute on a not-yet-delivered building turns a $14B valuation into a financing-heavy shell. The market is paying today for 2028 delivery.
2. Bitcoin crash + pledged collateral. ~$688M of BTC is pledged with margin-call/liquidation triggers (130%/105%); a sharp BTC drawdown forces sales into weakness, craters the GAAP P&L, and could reopen the equity-dilution spigot.
3. Governance overhang. A 2024 pump-and-dump short report + CEO exit, a Trump-family-front ABTC subsidiary (Lens 13), aggressive SBC, and consolidation optics that flatter headline numbers — any flare-up re-prices the "trust" component hard (HUT has shown it can drop 23% in a day on a report).
4. Sector funding gap. VanEck (Jun 16 2026) flags a ~$50B funding gap across the miner-to-AI cohort — if capital markets tighten, the IG-bond template that is HUT's edge could close.

Pre-mortem (18 months out, thesis broke): most likely cause — a River Bend or Beacon Point delivery slip into 2028 + a BTC drawdown hitting simultaneously; the stock had capitalized 2028 NOI at a REIT multiple, the cash didn't arrive on schedule, BTC marks gutted GAAP, and a confidence wobble (another short report, or an ABTC/Trump headline) compressed the multiple back toward pure-play-miner (5.9× sales). Secondary cause: a tenant credit event or renegotiation on the unnamed Beacon Point tenant.

Are multiples too high? On reported earnings, absurdly (GAAP loss). On 2028 contracted NOI, defensible-to-rich. The valuation requires 2027 execution; there is little margin of safety for slippage.

Contrarian view (what the market is refusing to see): The bears still frame HUT as "an inefficient bitcoin miner with a Trump-meme subsidiary." The filings say it has already become a uniquely-financeable hyperscale landlord with Google-grade credit behind its biggest lease — the bitcoin is now the option, not the thesis. The market is debating the wrong company. (The symmetric contrarian-bear: it's also refusing to fully price how much of that $14B is unearned 2028 cash flow that hasn't survived first contact with a construction schedule.)

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case:

• The whole equity is a forward bet on two buildings that don't exist yet. Strip the contracted backlog (unearned until 2027) and you have a GAAP-lossmaking bitcoin miner trading at ~$14B with 15,679 BTC (~$1.4B) and ~$405M corporate debt. The "platform" revenue is eliminated in consolidation — you literally cannot see it in the income statement yet.
• Revenue concentration is extreme and lopsided in time: ~93% of today's revenue is bitcoin mining (commodity, ABTC, ~80%-owned), while ~100% of the future equity story is two tenants (Fluidstack/Anthropic and one unnamed AA- name). Lose, delay, or renegotiate either lease and the thesis halves.
• The moat may be weaker than bulls think: every miner-peer is racing into AI data centers (IREN/Microsoft, Riot/Corsicana, TeraWulf, Core Scientific). Power/land is scarce but HUT is not the only one with a pipeline, and hyperscalers can dictate terms across many landlords. The IG-bond "first" is replicable now that the template exists.
• The most dangerous competitor bulls underestimate: IREN (already ~$18B with a live Microsoft contract at ~85% project EBITDA margin ) — it's ahead on actually-paying AI revenue.
• Worst governance/cap-allocation marks: the 2024 pump-and-dump short report + CEO departure; $50.9M/quarter SBC; an ABTC subsidiary fronted by Eric Trump (Chief Strategy Officer) and Donald Trump Jr. — a politically-charged, meme-adjacent structure that invites scrutiny, conflict-of-interest questions, and headline risk, and whose minority (~20%) leaks economics out of HUT shareholders' hands.
• What must hold for today's price: BTC doesn't crash (pledged-collateral triggers), both leases deliver on schedule in 2027, project debt stays serviceable, capital markets stay open, and no governance shock detonates. If growth/delivery disappoints 20–30% (a one-year slip + partial tenant haircut), a name capitalizing 2028 NOI at a REIT multiple re-rates toward pure-play-miner sales multiples — plausibly −40-60% downside.
• Single scenario that permanently impairs: a tenant credit/withdrawal event on Beacon Point combined with a BTC bear market — non-recourse project debt protects the parent legally but the equity narrative (and the cost-of-capital flywheel) would break. Plausibility: low-but-non-trivial.

Lens 14 · Management Questions (ordered by information value)

1. Beacon Point tenant: who is it, what is the exact credit rating, and what are the termination / non-delivery / renewal-decline provisions? (The single biggest unknown driving ~half the equity value.)
2. What is the construction critical path to Q2-2027 (River Bend) and Q3-2027 (Beacon Point) delivery, and what are the realistic slip scenarios on power interconnect and equipment (Vertiv/GPU/transformer lead times)?
3. Post-$7.5B of project bonds, what is the consolidated and corporate-only (recourse) debt-maturity ladder and DSCR, and the covenant headroom on the River Bend/Beacon Point notes?
4. On the ~$688M of pledged bitcoin: at what BTC price are margin calls/liquidations triggered across FalconX/Bitmain, and what is the liquidity plan if BTC falls 40% while a build is mid-construction?
5. What third-party (non-eliminated) Digital Infrastructure / Power revenue will be recognized in FY2027 vs FY2028 as the leases commence — give the ramp.
6. Why retain a bitcoin treasury at all now that the model is contracted infrastructure — does the BTC volatility raise your equity cost of capital more than the optionality is worth?
7. ABTC: what is HUT's exact economic % and voting %, the go-forward intercompany fee economics, and the plan for the NCI — buy-in, spin, or hold?
8. How do you think about further equity dilution (ATM + $50.9M/quarter SBC) given the explicit goal of funding builds with non-recourse debt — what's the dilution ceiling?
9. What is the stabilized, full-ramp consolidated NOI and adjusted-EBITDA bridge for FY2028 once both leases are paying, attributable to HUT shareholders (i.e. net of ABTC NCI)?
10. Beyond River Bend/Beacon Point, what is the conversion timeline for the 1,755 MW under exclusivity / 1,230 MW under development into signed tenant leases, and who are the prospective counterparties?
11. What is the Anthropic relationship beyond Fluidstack/River Bend — is there a path to a direct Anthropic offtake across the optional 2,295 MW?
12. How do you mitigate tariff/Bitmain single-source risk on miners and the equipment supply chain for the data-center builds?
13. What governance changes have been made since the 2024 leadership transition to address the concerns raised in the short-seller report?
14. What is the capital-return philosophy once contracted NOI is flowing — debt paydown, buyback, or continued growth capex?
15. Under what condition would you monetize bitcoin holdings to fund the platform rather than issue equity or debt?

---