Hut 8 Corp. (NASDAQ: HUT) — A Bitcoin Miner Reborn as a Power Landlord, Priced for the Whole Grid

Independent equity research. Report date: 2026-06-13.

⚡ Claude’s Take

This block is the author’s own independent opinion. It is general information only and not investment advice. Everything below it (the analysis) is deliberately position-free and carries no price target.

Verdict: AVOID at $118.86 / accumulate-on-weakness only / NOT a short. Fair-value zone ~$70–95 per share on the contracted lease book plus a modest pipeline credit and the ABTC/Bitcoin stub; the stock already pays for a fully-converted gigawatt pipeline, an investment-grade refinancing, and a benign Bitcoin tape — all at once. Conviction: medium.

Hut 8 has executed one of the cleaner pivots in the converted-miner cohort. Under Asher Genoot (CEO since February 2024) it has rotated from a commoditized, halving-squeezed Bitcoin miner into a “power-first” energy-infrastructure platform with two signed 15-year, triple-net, take-or-pay AI data-center leases totaling ~597 MW of critical IT and ~$16.8B of base-term contracted value — one anchored by an unnamed, disclosed AA-or-higher multi-trillion-dollar hyperscaler (Beacon Point, Texas), the other by a Google-backstopped FluidStack/Anthropic arrangement (River Bend, Louisiana). It financed them with ~$7.5B of oversubscribed, investment-grade-marketed, bankruptcy-remote project bonds non-recourse to the parent — institutional validation no slide deck can fake. It also carved its mining into the separately-listed American Bitcoin (ABTC), retaining ~55% economics, and is deliberately winding down its on-balance-sheet Bitcoin. The GAAP losses that scare retail (−$226M in FY2025) are a Bitcoin-mark sign-flip, not a cash bleed. This is a real, well-built business.

But the price already discounts near-perfection. The stock has roughly tripled off its FY2025 close ($45.94) to $118.86, sitting at the 97.9th percentile of its own ~10-year valuation history (P/S 99.6th, P/B 94.9th) and the richest EV-per-contracted-MW (~$30M/MW) in the entire converted-miner landlord group (vs. ~$15–25M for CIFR/WULF, ~$12–18M for APLD) — even though not a dollar of HPC rent will be recognized until ~mid-2027. Capitalizing management’s ~$1.1B stabilized NOI at a fair ~7% cap, netting the ~$7.5B of project debt, and adding the ABTC stake (~$0.5B) and the shrinking Bitcoin stub (~$0.6B) gets you to roughly $70–95 a share; the gap to $118.86 is the market pre-paying for the unsigned ~1 GW+ pipeline and an IG refi that has not happened. On top of that sits a 2.4-beta crypto tail (this same security fell ~95% in the last crypto winter) and a Trump-family-affiliated subsidiary carrying genuine related-party and headline risk. The framing is “great contracted book, momentum price, crypto-beta tail” — a crowded, high-beta one-way-street-up trade sold at its richest-ever multiple, mid-capital-cycle. I would not short it (14–16% of the float is already short, every covering analyst rates it Buy, the catalysts are live, and the squeeze/Bitcoin-melt-up risk is acute), but I would not pay up here. Flips bullish: a meaningful drawdown into the ~$70s, OR an investment-grade refinancing of the SPV notes that proves a levered equity spread, OR a second hyperscaler signing on the remaining Beacon Point capacity. Flips bearish: a slipped 2027 lease commencement against the cash-interest wall, a Bitcoin crash that drags the 2.4-beta equity, or any crack in the FluidStack/Anthropic neocloud credit. Tag: “They built the substation beautifully — then sold you the whole power grid at retail.”

1. Executive Summary

Hut 8 Corp. (NASDAQ/TSX: HUT) is a Miami-headquartered, vertically-integrated energy-infrastructure and digital-compute platform — until recently one of North America’s larger Bitcoin miners, now in a fast, deliberate metamorphosis into a landlord that builds and leases hyperscale AI/HPC data centers to investment-grade tenants under long-dated, triple-net contracts. The transformation has been led by CEO Asher Genoot since February 2024 and reorganized the company into three reportable segments — Power (power origination, development, generation, managed services), Digital Infrastructure (build-own-operate data centers — the AI-landlord engine), and Compute (Bitcoin self-mining via the majority-owned American Bitcoin, plus cloud/GPU services).

The investment debate is unusually clean and unusually prospective. What is real: two signed campus leases — Beacon Point (Nueces County, TX; phase 1 of a planned 1 GW campus): 352 MW critical IT, 15-year term, ~$9.8B base-term value (up to ~$25.1B with renewals), leased to an unnamed but disclosed AA-or-higher multi-trillion-dollar technology company; and River Bend (West Feliciana Parish, LA): 245 MW IT, 15-year term, ~$7.0B base-term value, leased to FluidStack with a Google financial backstop and Anthropic as the ultimate compute user. Together: ~597 MW of contracted IT, ~$16.8B of aggregate base-term lease value, and ~$1.1B of stabilized annual net operating income (NOI) — but commencing only around Q2–Q3 2027. Hut 8 financed the build with ~$7.5B of project-level senior secured notes (River Bend: $3.25B at 6.192%; Beacon Point: $4.25B at 6.129%), each issued by a bankruptcy-remote subsidiary, IG-marketed, and non-recourse to the parent. It also owns ~55% of the separately-listed American Bitcoin (ABTC) and a Bitcoin treasury it is actively shrinking.

What is unproven: as of Q1-2026, zero HPC lease revenue has been recognized — all $235M of FY2025 revenue is Compute (Bitcoin mining + hosting + cloud), recognized largely as Bitcoin-in-kind, not cash. The company is deeply free-cash-flow-negative (~−$382M FY2025) and has never self-funded; growth has been financed by serial low-price equity issuance (now pivoting to the project bonds). GAAP earnings are dominated by non-cash Bitcoin fair-value remeasurement — a +$509M gain in FY2024 flipped to a −$220M loss in FY2025, which by itself swamps the entire net-income swing. Insiders have never bought a share on the open market, directors sold into the parabolic run, and the equity carries a 2.4-beta link to Bitcoin and a Trump-family-affiliated subsidiary.

Valuation is the crux. HUT trades at the 97.9th percentile of its own ~10-year P/S and P/B history, at ~$30M EV per contracted MW — the richest in its cohort — and at an implied capitalization of management’s not-yet-earned NOI that sits at or below the cost of its own project debt. That leaves essentially all equity value at $118.86 resting on the uncontracted pipeline converting, an eventual IG refinancing, and a benign Bitcoin tape. This memo takes no position; it lays out the embedded expectations and the falsification tests. The single, fenced-off opinion is in Claude’s Take above.

2. Business Overview

What the company does. Hut 8 conceives, builds, finances, and operates the physical stack that energy-intensive computing requires: it originates power and grid interconnection, develops and constructs data centers, and operates them — for its own Bitcoin mining, for hosting third parties, and, increasingly, as a landlord delivering powered, built-out capacity to hyperscale AI tenants. The corporate identity, restated through 2025–2026, is an “energy infrastructure platform integrating power, digital infrastructure, and compute at scale.” The company was founded in 2017, is headquartered in Miami, employs ~222 people, reports on a December fiscal year (it changed its year-end from June to December, hence the FY2023 transition-period filing), and is dual-listed on Nasdaq and the TSX.

The business is a barbell of a declining/commoditized legacy and a ramping, contracted future, organized into three segments:

• Compute (the legacy + the captive miner). This is essentially American Bitcoin Corp (ABTC) — the entity into which Hut 8 contributed its ASIC fleet in 2025, now separately listed (it went public via a reverse merger with Gryphon Digital Mining in September 2025) and consolidated by Hut 8 at ~55% economic ownership. ABTC is a “pure-play Bitcoin accumulation” vehicle (its headline metric is Bitcoin-per-share); Hut 8 is effectively its landlord/host, earning ~20–25% yield-on-cost hosting economics while ABTC self-funds its own chip capex. The segment also includes traditional cloud, equipment sales/repair, and the Highrise AI neocloud (a wholly-owned, GPU-owning AI-cloud business of ~1,100 GPUs scaling toward ~20,000 — the one place Hut 8 owns the compute). Compute is ~86% of FY2025 revenue but recognized largely as Bitcoin-in-kind.

• Digital Infrastructure (the new core). The build-own-operate data-center engine: the two signed campuses (Beacon Point, River Bend) and the development pipeline. This segment generates no revenue yet — the first data hall energizes ~Q2–Q3 2027 — but it is where management argues the durable value lives, and it is what the equity is priced on.

• Power (the origination edge). Power generation, energy-portfolio optimization, and managed services (site design, procurement, construction management, hosting operations) sold to third-party site owners, governments, and developers. This segment housed the four Ontario natural-gas plants (310 MW, a Macquarie JV) that Hut 8 sold to TransAlta in early 2026, booking a one-time gain — a portfolio-optimization move that simultaneously shrank reported Power revenue ~59%.

American Bitcoin (ABTC) deserves its own paragraph, because it is both a value driver and the single biggest source of structural complexity. In 2025 Hut 8 contributed essentially its entire ASIC mining fleet into a vehicle (originally “American Data Centers”) renamed American Bitcoin, in which it took ~80% of the economics and the Trump-family founders (Eric Trump — a co-founder and on management/board — and Donald Trump Jr. among the founders) took the balance. ABTC then went public via a reverse merger with the listed shell Gryphon Digital Mining (closed September 2025), giving it its own ticker, its own ATM (~$237.7M) and a PIPE (~$205.3M), and its own “pure-play Bitcoin accumulation” mandate (its headline KPI is Bitcoin-per-share). The consequences for Hut 8’s financials are large and easy to misread: (1) Hut 8 consolidates ABTC, so 100% of ABTC’s revenue, Bitcoin holdings, and — critically — Bitcoin fair-value marks flow through Hut 8’s income statement and balance sheet; (2) but as ABTC issued its own equity, Hut 8’s economic stake fell from ~80% to ~55%, so ~45% of ABTC’s results are stripped back out as noncontrolling interest (the $310M NCI line and the ~$21.8M of FY2025 net loss attributed to minorities); and (3) Hut 8 is simultaneously ABTC’s landlord/host, earning ~20–25% yield-on-cost hosting economics from it — so ABTC is both a consolidated subsidiary and a captive customer. For valuation, the right treatment is to value Hut 8’s ~55% stake at ABTC’s (currently depressed) public market price and not to double-count the Bitcoin that sits inside ABTC against the minority.

How it makes money — today vs. tomorrow. Today, revenue is overwhelmingly Bitcoin-economics: block rewards and hosting/yield from ABTC, recognized as non-cash Bitcoin consideration whose value swings with the coin price. Tomorrow, the model flips to contracted triple-net rent — long-dated leases under which Hut 8 delivers powered, built-out data-center capacity and the tenant bears essentially all operating cost (>99.9% of rent is projected to drop to NOI). That converts a commodity-price-taking cash flow into durable, investment-grade-counterparty contractual income. The recurring-revenue quality therefore improves dramatically on paper — but the recurring stream has not started.

Power “under management” was ~1,020 MW across 15 sites at year-end 2025 (including the since-sold Ontario plants); operating behind-the-meter and hosting sites include King Mountain (280 MW wind, a NextEra JV), Vega (205 MW), and Granbury (300 MW). The pipeline of grid capacity in development/exclusivity is far larger — management frames a platform “exceeding 2.5 GW under management” and an interconnection pipeline of multiple gigawatts (including a 1 GW ERCOT position near Corpus Christi that underlies Beacon Point). The “>2.5 GW” figure is developed-plus-contracted-plus-pipeline ambition, not operating capacity — a distinction that matters for valuation.

The contract architecture matters more than the headline MW. Each campus is housed in a bankruptcy-remote project subsidiary (Hut 8 DC LLC for River Bend; Beacon Point DC LLC for Beacon Point), which both isolates risk and is the unit against which the project bonds are issued. The “company” is therefore best understood as a holding entity sitting atop (a) a small portfolio of single-asset, separately-financed data-center projects, (b) a ~55% stake in a separately-listed Bitcoin miner, © a shrinking Bitcoin treasury, and (d) a power-origination pipeline. Verdict: a coherent, well-articulated transition from price-taking commodity production to contracted infrastructure rent — but as of this report the new model is two signed contracts and two construction sites, not yet a revenue line.

3. Industry Dynamics

The demand backdrop is as strong as any in the market. Hyperscaler AI capital expenditure is running at a ~$700B/year pace across the largest cloud and model builders, and the binding constraint has shifted decisively from chips to power and the data centers to house it. Interconnection queues, multi-year transformer and switchgear lead times, and local grid limits mean that whoever controls energized (or near-energized) large-block power at speed controls a scarce, monetizable asset. The “power is the bottleneck” thesis — well established across prior published work on CIFR, APLD, WULF, IREN, OKLO, VST, and CEG — is, in our assessment, correct and durable for at least the next several years. Hut 8’s deliberate siting in ERCOT (Beacon Point, on AEP Texas) and Louisiana (River Bend, on Entergy) gives it faster/cheaper interconnection than the notoriously clogged PJM queue, plus grid and geographic diversification that multi-region hyperscalers value.

Supply is where the caution lives. The economics are visible enough (premium colocation rents, high-80s%-to->99% NOI margins on stabilized triple-net assets, asset-level IRRs quoted in the high-teens-to-20s%) that capital is flooding in from every direction: the converted Bitcoin miners (CIFR, WULF, IREN, APLD, CORZ, RIOT, MARA), purpose-built neoclouds (CoreWeave, Crusoe, Nebius, FluidStack), the established colocation REITs and privates (Digital Realty, Equinix, Vantage, QTS, Switch), PE infrastructure vehicles, and — most importantly — the hyperscalers self-building. The hyperscaler is simultaneously Hut 8’s best customer and its largest potential competitor. In Marathon capital-cycle terms, the industry sits mid-to-late cycle: demonstrably high returns are pulling in a wave of capital that will, on a multi-year horizon, compress rents and returns. Current premium rents and >99% triple-net NOI margins are best understood as cyclical-high economics, not a permanent plateau.

The Bitcoin-mining leg is a structurally bad industry, and Hut 8 knows it. Bitcoin mining is fully commoditized — output is a homogeneous global commodity, there is no pricing power, and the protocol’s periodic halvings (the most recent in April 2024) mechanically compress “hashprice” (revenue per unit of hash) over time, forcing perpetual reinvestment in newer, more efficient ASICs just to stand still. This is the textbook Marathon “bad industry”: high capital intensity, no demand captivity, returns that mean-revert toward the cost of (cheap) power. Hut 8’s strategic response — push the mining into a separately-funded vehicle (ABTC) and redeploy its scarce asset (power + land + interconnection) into contracted AI landlord economics — is the rational move, and is precisely the cohort-wide pivot the capital-cycle lens would predict once mining returns compressed.

The capital-cycle lens (Marathon) is the single most important industry frame, and it cuts against the equity even as it validates the strategy. The Marathon discipline says: follow the supply of capital, not the demand. When returns in an industry are visibly high, capital floods in on the supply side, capacity overshoots, and returns mean-revert — and returns look best precisely at the top, when everyone can see them and the capital is still arriving. Mid-2026 is a textbook expression of that pattern in AI data-center development: converted miners, neoclouds, REITs, sovereign and PE infrastructure vehicles, and the hyperscalers themselves are all racing to lock up power and build, debt and equity capital is being raised in the tens of billions, and the marginal economics (rent/kW, NOI margin, asset IRR) are being quoted at levels that assume the scarcity persists. History — telecom fiber ~2000, dry-bulk shipping ~2007, US shale ~2014, the prior crypto-mining boom ~2021 (which took this very security down ~95%) — says it usually does not. The signal to watch is not demand (which is real and may stay strong) but the spread between new-build cost and achievable rent, and the cap rate the market pays for contracted NOI. When that cap rate compresses below the cost of the debt funding the build — as it arguably has for HUT’s whole-EV — the cycle is pricing perfection. None of this means HUT’s signed projects are bad; asset-level returns on already-leased, IG-anchored capacity can be excellent. It means the industry-wide return on the next wave of capital — the pipeline the equity capitalizes — is more likely to disappoint than to surprise.

Regulatory and power structure. ERCOT is an energy-only, weather-volatile grid (the Winter Storm Uri precedent), exposed to curtailment, basis risk, and an evolving large-load interconnection regime (PUCT/ERCOT rules for multi-hundred-MW loads are still being finalized) — a live source of timeline uncertainty for the un-energized pipeline. Louisiana/Entergy (an integrated, regulated utility) offers a different, more stable model. Large-load data-center economics are also drawing political attention (ratepayer cost-shift concerns), a slow-moving regulatory risk for the whole sector.

Barriers to entry are real but replicable: long-lead power origination, interconnection rights, speed-to-build execution, and access to project capital. None is a structural monopoly; each is a head-start a well-funded competitor can close in 18–36 months. Verdict: a structurally attractive industry on the demand side and a structurally crowded, late-capital-cycle one on the supply side — overlaid on a structurally bad (commoditized, halving-squeezed) mining business that Hut 8 is correctly de-emphasizing. It is a good industry for an incumbent that already controls energized power and signed IG tenants, and a more dangerous one for the fresh capital — including the pipeline value embedded in HUT’s equity — deployed at mid-2026 valuations.

4. Competitive Position

Name the moat — or its absence. Applying Greenwald’s taxonomy, Hut 8 does not possess a durable competitive advantage. It possesses a head start, a top-tier tenant-credit anchor, and an unusually diversified (but messier) asset mix.

• Cost / supply advantage (partial, replicable): in-house power origination and a multi-gigawatt interconnection pipeline (the 1 GW ERCOT position near Corpus Christi, the Louisiana/Entergy greenfield, behind-the-meter wind at King Mountain) is a genuine access edge assembled over years. But interconnection rights and construction speed are acquirable — there is no proprietary technology, patent, or regulatory franchise locking competitors out.
• Customer captivity (contractual, post-signing only): once a 15-year triple-net lease is signed with an IG hyperscaler, switching costs are high and the cash flow is sticky — but this captivity exists only on the ~597 MW already contracted. The ~1 GW+ pipeline that the market is capitalizing has zero captivity; every megawatt must still win a competitive tenant negotiation. ABTC provides a measure of captive internal demand (Hut 8 hosts its own ~55%-owned miner), which usefully de-risks the early ramp of greenfield power but is not a third-party moat.
• Economies of scale / network effects: absent. At ~597 MW contracted, HUT is sub-scale versus the colocation REITs and has no network effect (single-tenant hyperscale colocation is not a two-sided network).

Where HUT genuinely differentiates — and it matters — is tenant credit on Beacon Point. The Beacon Point anchor is an unnamed but disclosed AA-or-higher, multi-trillion-dollar-market-cap technology company (management has confirmed it is not Anthropic, Google, or FluidStack — i.e., a genuine hyperscaler of the Microsoft/Amazon/Alphabet/Meta/Apple/NVIDIA class). A disclosed-rating, direct hyperscaler anchor is arguably the single highest-quality counterparty in the converted-miner cohort — cleaner even than CIFR’s unnamed third tenant, and materially stronger than APLD’s reliance on CoreWeave (a ~BB-standalone, Microsoft-concentrated neocloud). River Bend is the lower-credit leg: FluidStack is a neocloud whose lease required a Google financial backstop (Google backstops base-term rent and pass-throughs), with Anthropic as the ultimate compute user — the same FluidStack/Google/Anthropic credit cluster that anchors WULF’s Lake Mariner and CIFR’s Barber Lake. So Hut 8’s contracted book pairs one best-in-cohort direct hyperscaler with one credit-wrapped neocloud — a sensible barbell, but one that still leaves the company a three-counterparty-concentrated developer (hyperscaler + FluidStack/Google + ABTC) rather than a diversified REIT.

The landlord-vs-neocloud distinction — the most important risk fork in the cohort — places HUT mostly on the safer side. A neocloud (CoreWeave, IREN, Crusoe) buys the GPUs — tens of thousands of accelerators that depreciate on a 3–6-year curve, face obsolescence with each NVIDIA generation, and must be kept utilized to service the debt that bought them. A landlord (HUT, CIFR, WULF, APLD) builds the shell and the power and lets the tenant bear the GPU risk; the asset is the building, the interconnection, and the lease — far longer-lived and not subject to Moore’s-law obsolescence. Hut 8’s two signed campuses are pure triple-net landlord deals (tenant owns the GPUs). The one exception is Highrise AI, a wholly-owned, GPU-owning neocloud — small today (~1,100 GPUs) but a deliberate toe into the higher-risk, higher-return owned-compute model. The flip side of the landlord posture: HUT captures less upside if AI compute economics stay euphoric, and its terminal value still depends on the buildings being re-leasable after the initial 15-year terms.

Run the Greenwald tests explicitly. Market-share stability: there is no stable share to measure — this is a new, fast-growing, fragmenting market where positions are being established by who signs leases this year, the opposite of the stable-share signature of a real moat. ROIC test: not yet measurable on the HPC business (pre-revenue); the legacy mining ROIC was unremarkable and structurally mean-reverting. Barriers-to-entry test: the barriers (power, interconnection, capital, speed) are access advantages a determined, well-capitalized entrant can replicate, not the demand-captivity or proprietary-cost barriers Greenwald requires. HUT passes none of the three cleanly. Verdict: a credible head start with the best direct hyperscaler anchor in the converted-miner cohort, the lower-risk landlord posture, and captive ABTC demand — but a replicable land-grab, not a durable franchise. Counterparty concentration, an unnamed (if rated) anchor, a credit-wrapped neocloud, and unproven 2040s terminal values are real fragilities the market is not currently pricing.

5. Growth History and Forward Opportunities

History. Revenue grew from ~$79M (FY2022) to $162.4M (FY2024) to $235.1M (FY2025), and Q1-2026 revenue rose to $71.0M from $21.8M a year earlier (+226%) — but this growth is Bitcoin-economics growth (more hashrate, more hosting, plus the BTC price), recognized largely as Bitcoin-in-kind rather than cash, and it is now being deliberately de-emphasized. The reported top line is therefore an unreliable guide to either the past or the future: it overstates cash generation (the non-cash “ASIC compute revenue” add-back was ~$187M in FY2025), and it will be increasingly dominated by the contracted-rent story that has not yet begun.

The forward opportunity is the entire thesis. Management’s contracted book is the central artifact:

• The two signed leases: ~597 MW of IT, ~$16.8B aggregate base-term value over the initial 15-year terms (Beacon Point ~$9.8B / River Bend ~$7.0B), rising to >$25B (Beacon Point alone, with three 5-year renewals), at ~$1.1B of stabilized annual NOI (~$655M from Beacon Point), with 3% annual escalators and triple-net, take-or-pay structures under which >99.9% of rent drops to NOI. First data halls energize ~Q2–Q3 2027.
• Beacon Point campus expansion: phase 1 (352 MW) is the first slice of a planned 1 GW campus — leaving ~650 MW of additional capacity at an already-permitted, already-financed, NVIDIA-DSX-architected site to lease to a second tenant.
• River Bend expansion: the 245 MW phase 1 sits within a planned 1 GW campus, and FluidStack holds a right-of-first-offer on an additional 1,000+ MW — a multi-gigawatt partnership option.
• The broader power pipeline (multi-GW): additional ERCOT and other-grid interconnection positions in development, the platform “exceeding 2.5 GW under management” ambition, and behind-the-meter generation optionality (West Texas gas).
• ABTC + Bitcoin treasury: a ~55% stake in a public, ATM-funded Bitcoin accumulator, plus a (shrinking) corporate Bitcoin treasury — crypto-beta optionality that can compound or crater with the coin.
• Highrise AI: the owned-GPU neocloud scaling toward ~20,000 GPUs — a higher-risk, higher-return adjacency.

Unit economics, as far as they can be reverse-engineered. ~$1.1B of stabilized NOI on ~597 MW of IT implies roughly ~$1.8M of annual NOI per critical-IT MW (~$150/kW/month-equivalent) — at the premium end of the colocation economics observed across comparable names, consistent with the scarcity of energized large-block power in 2026 and the take-or-pay/triple-net structure. The bull reads this as proof the model compounds: each delivered site de-risks and helps finance the next, and the AA-rated anchor proves Hut 8 can win the best counterparties. The skeptic notes that premium pricing is itself a cyclical-high phenomenon, that the NOI is a portfolio target hiding per-asset spreads and tenant-specific terms not in the filings, and that ~$1.1B of NOI capitalized at any reasonable rate is already inside the EV — so the growth that matters for the equity is the uncontracted pipeline, which is unsigned.

The owned-compute and behind-the-meter options are real but unproven. Two adjacencies sit outside the pure-landlord model and could swing the return profile in either direction. Highrise AI is a wholly-owned neocloud — Hut 8 buys and owns the GPUs (~1,100 today, scaling toward a stated ~20,000) and sells bare-metal and cloud compute. If AI compute pricing stays euphoric, an owned-GPU book earns far more than triple-net rent; if utilization or pricing slips, it becomes the capital-destruction machine that the landlord model was designed to avoid (the IREN/CoreWeave risk). It is small enough today to be optionality rather than thesis, but it signals that management’s “we prefer to be the landlord” discipline is not absolute. Behind-the-meter generation is the other option: West Texas sites sitting atop cheap natural gas could, in principle, be powered off-grid, bypassing interconnection queues entirely — a potential source of gigawatts of fast capacity and a structural cost edge. It is also, like everywhere in the cohort, an unsolved engineering/permitting/financing challenge with nothing in the numbers. Both are free options layered on the contracted book; neither should be paid for as if delivered.

Quality of growth. If delivered, this is high-quality growth: long-duration, contracted, investment-grade-counterparty, triple-net cash flow replacing volatile commodity production. But the quality is entirely prospective. The growth is capital-intensive (FCF deeply negative through commencement), debt-and-equity-funded, and gated by construction execution, interconnection approvals partly outside the company’s control, and tenant negotiations not yet won on the pipeline. And the reported revenue mix will remain Bitcoin-dominated and volatile until the leases commence in 2027. Verdict: potentially very high-quality contracted growth, but unproven, front-loaded with execution and funding risk, and not yet earning a dollar — the single largest gap between narrative and realized cash flow in this report.

6. Financial Quality

The headline loss is mostly a Bitcoin-mark sign-flip — read past it, but don’t mistake it for strength. Hut 8’s GAAP result swung from +$331.9M net income (FY2024) to −$226.1M net loss (FY2025), a reversal that looks alarming until you decompose it. The dominant driver is the “loss (gain) on digital assets” line under fair-value accounting (ASU 2023-08): a +$509.3M gain in FY2024 flipped to a −$220.0M loss in FY2025 — a $729M swing that by itself exceeds the entire net-income swing. Layer on non-cash derivative/warrant remeasurement (FY2025 ~+$61.6M, largely the Coatue convertible’s embedded conversion option) and a one-time Q1-2026 TransAlta divestiture gain (~+$33.6M). The correct conclusion: do not capitalize the GAAP loss — but do not read its non-cash nature as financial health either.

The single most important row is the first non-net-income line: the Bitcoin fair-value mark alone swung ~$729M between the two years — larger than the entire net-income swing — which means the headline result tells you almost nothing about the operating business and almost everything about the coin price. The other rows reinforce two themes the rest of this section develops: a chunk of the “revenue” is non-cash Bitcoin-in-kind (so the income statement overstates cash), and the one genuinely recurring adjustment management makes (SBC) is the one it adds back.

Management’s “Adjusted EBITDA” is an illusion you must reject. Critically, the company’s own Adjusted EBITDA reconciliation does not strip the Bitcoin mark — so FY2024 “Adjusted EBITDA” of ~$555.7M was manufactured by leaving the $509M gain in, and FY2025 was ~−$135.4M with the $220M loss in. Worse, the metric does add back stock-based compensation, which ballooned to $57.8M in FY2025 (from $20.8M in FY2024, +178%) and ~$50.9M in Q1-2026 alone. Adjusted EBITDA that includes unrealized crypto marks and excludes real, growing dilution is doubly misleading — and (see ) it is the metric management is paid on.

Cash flow tells the true story: the business does not self-fund. Operating cash flow was −$68.5M (FY2024) and −$139.2M (FY2025) (−$27.2M in Q1-2026). FY2025 investing was −$754.2M (including ~$405M of ABTC Bitcoin purchases, ~$203M of PP&E, and ~$163M of land deposits), and financing was +$856.1M — essentially 100% external (debt + equity). Free cash flow was roughly −$382M in FY2025. The headline revenue growth materially overstates cash generation because mining revenue is booked as Bitcoin-in-kind (the ~$187M FY2025 non-cash add-back). This is a cash-consuming infrastructure build, financed entirely from capital markets.

Balance sheet — the heart of the matter. At Q1-2026 (before the June Beacon Point bond): total assets $2.61B; unrestricted cash $160M; Bitcoin treasury ~15,679 coins, ~$1.37B gross (split across custody ~$436M + pledged-for-miners ~$211M + pledged-as-collateral ~$478M — i.e., much of it encumbered); PP&E net $812M; goodwill $209M (largely from the Gryphon/ABTC reverse merger); total equity $1.69B, of which $310M is noncontrolling interest (the ~44.7% of ABTC owned by third parties, including the Trump-affiliated founders). Net debt was only ~$245M at that date — but that figure is now obsolete.

The debt architecture — the strength and the risk. In a six-week span Hut 8 issued two non-recourse project bonds totaling ~$7.5B:

Both are issued by bankruptcy-remote subsidiaries, IG-marketed and oversubscribed/at-par, escrowed for construction plus debt-service reserves, amortizing from ~2030, and non-recourse to the Hut 8 parent. The genius of the structure is that a single project’s failure cannot drag down the whole company, and the construction risk is ring-fenced; the River Bend close even recovered ~$184M of equity to the parent at ~95% loan-to-cost. The corresponding risk: the escrows deplete as construction proceeds, and once each project is built the bond’s ~$230M+/year (combined) cash interest must be serviced from that project’s lease NOI — which does not begin until ~mid-2027. The window between escrow depletion / construction completion and stabilized lease revenue is the precise financial pinch-point. The equity-spread thesis ultimately depends on refinancing these 6.1–6.2% notes toward investment grade once the assets stabilize — not yet achieved.

Parent-level debt is small and shrinking. Pre-bond corporate debt was modest (~$410M): a TZRC note (~$49.6M), a $200M Coinbase credit facility at 9.0% maturing June 2026 (near-term refi/repayment item), an undrawn $200M revolver, and the Coatue convertible (~$159M face, 8.0%, conversion $16.395) — now deeply in-the-money at $118.86, carried at $159M but fair-valued near $459M (~$300M of unrecognized economic dilution, ~9.7M shares). Management expects the Coatue convert to force-convert around late June 2026, after which it asserts near-zero parent-recourse debt and ~$1.3B of parent liquidity. So the consolidated balance sheet carries ~$7.5B of debt, but it is almost entirely project-level and non-recourse; the parent itself is lightly levered — a genuinely important and favorable structural distinction versus a naive “$7.5B of debt” read.

Dilution and SBC. Shares grew from ~99.5M (Dec-2024) to ~110.1M (Dec-2025) to ~112.5M (Q1-2026), funded by serial ATM issuance at low prices (the weighted-average issuance price since commencement was only ~$18.61 — i.e., the company sold a great deal of stock far below today’s price). Q1-2026 alone added ~$120M of Hut 8 equity plus ~$110M of ABTC equity. The explicit pivot from dilutive equity toward ring-fenced bonds is the capital-structure story of the past year and a real maturation — but the dilution legacy (and the ITM convert) is a permanent feature of the per-share math.

Bitcoin treasury is being deliberately wound down at the parent (“it’s just an asset”), with crypto exposure retained through the ABTC stake; post-quarter the company refinanced its Bitcoin-backed borrowing from a 9% Coinbase facility into a ~7% FalconX 364-day note, freeing collateral. Verdict: economics will improve dramatically with scale if the leases commence and stabilize, but the present-day financial profile is a deeply FCF-negative, externally-funded build whose GAAP earnings are noise (Bitcoin marks) and whose “Adjusted EBITDA” is actively misleading. The non-recourse, parent-light debt structure and the at-par institutional bonds are real strengths; the equity quality is speculative until rent flows in 2027, and the security carries a Bitcoin-price sensitivity a pure data-center landlord would not.

7. Capital Allocation

A sophisticated, all-in, increasingly well-structured bet — undercut by weak alignment. Three threads:

(1) Financing strategy — genuinely impressive and improving. The rotation from a serially ATM-dilutive equity story (selling stock at a ~$18.61 average) to per-asset, non-recourse, investment-grade-marketed project bonds is the right architecture and a clear maturation: it isolates construction risk inside bankruptcy-remote SPVs, aligns debt service with each lease’s cash flows, and — critically — the bonds priced at par / oversubscribed, an unusually credible third-party endorsement of the contracts and the build plan. Raising ~$7.5B of IG-marketed paper against signed AA-/Google-backstopped leases, and recovering ~$184M of parent equity at the River Bend close, is best-in-cohort execution. The active portfolio management — selling the 310 MW Ontario gas plants to TransAlta (a ~$33.6M one-time gain and a cleaner power focus), refinancing Bitcoin-backed debt from 9% to ~7%, and carving mining into a separately-funded vehicle — all show a management team optimizing the capital structure rather than just issuing stock into every opportunity.

(2) The ABTC carve-out — clever monetization, but dilutive and governance-laden. Spinning the ASIC fleet into American Bitcoin and taking it public via the Gryphon reverse merger created a separately-funded vehicle that self-finances its own chip capex (removing a capital sink from Hut 8) and an ATM-funded “Bitcoin accumulation” currency. But Hut 8’s economic stake fell from ~80% at formation to ~55% as ABTC issued its own equity (a $237.7M ATM plus a $205.3M PIPE), so ~45% of ABTC’s economics — and its Bitcoin marks — now bleed to minorities (the $310M NCI). And ABTC is Trump-family-affiliated (Eric Trump a co-founder and on management/board; Donald Trump Jr. among the founders; the legacy “American Data Centers” vehicle), a structural related-party and political/headline exposure that has already produced negative press and post-listing share weakness. Capital-allocation-wise the carve-out is defensible; governance-wise it is a flag.

(3) Incentive design — the alignment is poor. The most damning capital-allocation finding is in the proxy: management’s long-term incentive (PSU) plan is keyed to “Adjusted EBITDA including Bitcoin fair value” — i.e., the very metric that bakes the non-cash, market-driven Bitcoin mark into the payout, rewarding management for a rising coin price rather than operational delivery. That is precisely backwards for a company trying to convince the market it is now an infrastructure landlord, not a Bitcoin proxy. The say-on-pay vote drew ~45% opposition — an extraordinary level of shareholder dissent that signals the market shares this concern. Combined with the SBC that ballooned 178% in a year, the compensation program is the weakest link in an otherwise-competent capital story.

The insider signal is unambiguously negative. Across the entire ~66-filing Form 4 corpus, there is not a single open-market purchase (code P) — no officer or director has bought a share at any price, even as the stock ran from the single digits to ~$119. Every acquisition is a $0 grant or option/RSU exercise; every disposition is a discretionary sale or tax withholding. Directors sold into the parabolic 2026 run (e.g., ~34K shares around $105–110 and ~20K around $101 in May 2026). CEO Genoot holds grants and his merger-related stake but has neither bought nor sold — PSU-locked, neutral. Verdict: a sophisticated, materially-improved financing program (non-recourse IG bonds, active portfolio management, parent deleveraging) — undercut by a Bitcoin-linked incentive metric, a ~45%-against say-on-pay, heavy SBC, the ABTC governance overhang, and a complete absence of insider conviction buying against directors selling the rally. Competent on the structure, poor on the alignment; the jury on durable per-share value creation is still out.

8. Changes and Headwinds — Last Two Years

The last ~24 months are the story — this is a company re-engineered in real time under a new CEO:

• New leadership & strategic pivot (Feb 2024 →): Asher Genoot became CEO and reframed Hut 8 from a Bitcoin miner into an “energy infrastructure platform”; the segment structure (Power / Digital Infrastructure / Compute) followed.
• American Bitcoin carve-out & IPO (2025): ASIC fleet contributed to ABTC (March 2025); public via the Gryphon Digital Mining reverse merger (September 2025); Hut 8’s stake diluted from ~80% to ~55% via ABTC’s own ATM/PIPE.
• Portfolio optimization (late 2025 – early 2026): the 310 MW Ontario gas-plant portfolio (Macquarie JV) sold to TransAlta (closed early 2026, ~$33.6M gain); Bitcoin-backed debt refinanced from 9% (Coinbase) toward ~7% (FalconX); DIFC license obtained for treasury/derivatives.
• The two anchor leases (2026): Beacon Point (352 MW, AA-rated hyperscaler, announced ~May 2026 — the stock jumped ~30–37% on signing) and River Bend (245 MW, FluidStack/Google/Anthropic) — together ~$16.8B base-term contracted value.
• ~$7.5B of project bonds (Apr–Jun 2026): $3.25B River Bend (6.192%) and $4.25B Beacon Point (6.129%), both non-recourse SPV notes — taking the company from near-debt-free to ~$7.5B of (ring-fenced) project debt; the stock traded lower on the June 5 Beacon Point pricing.
• Governance/IR: E. Stanley O’Neal (ex-Merrill Lynch CEO) appointed Board Chair (June 11, 2026); a new Head of Investor Relations hired from NextEra; the Coatue convert set to force-convert ~late June 2026.

Headwinds / watch-items: Bitcoin price (the 2.4-beta equity still moves with the coin, and the treasury/ABTC marks dominate GAAP); the 2027 lease-commencement timeline against the project-bond interest wall; ERCOT large-load interconnection-rule finalization; refinancing/rate risk on the 6.1–6.2% notes; the ABTC/Trump related-party and political headline risk; and the capital-cycle rent-compression risk on the uncontracted pipeline. Verdict: the changes overwhelmingly advance the thesis — they are the thesis — but each raises the stakes; the company is far less financially resilient to a Bitcoin or capital-markets shock than the lightly-levered miner it was two years ago, and the equity now embeds a demanding multi-year execution path.

9. Risk Analysis

Net: the dominant risks are Bitcoin price, 2027 commencement timing vs. the interest wall, tenant/pipeline conversion, refinancing, and dilution — a cluster that largely resolves over the next 12–24 months, making this an unusually datable thesis. A total loss is unlikely given the non-recourse ring-fencing, parent deleveraging, and at-par bonds, but a severe equity drawdown is entirely plausible in a Bitcoin-and-AI-narrative shock given the 2.4 beta and the rich entry valuation.

10. Valuation Discussion (Embedded Expectations)

HUT is expensive against its own history and priced for the full pipeline on its forward economics. Trailing multiples are distorted (GAAP is Bitcoin-mark noise; HPC revenue is zero), so the analysis works off own-history percentiles, an EV-per-MW cross-check, and a sum-of-the-parts — while flagging that the key NOI inputs are unaudited management metrics, not yet-filed lessor schedules.

Own-history screen — the single highest-signal datum. On AZI’s own-history valuation index (2026-06-12), HUT sits at the 97.9th percentile of its ~10-year history (composite): P/S 116x (99.6th percentile), P/B 22.0x (94.9th), P/E 76x (99.3rd — distorted by Bitcoin-mark-suppressed TTM EPS, so lean on P/S and P/B). The stock has roughly tripled off its FY2025 close of $45.94. By its own multi-year standard, this is the richest the equity has ever been — the inverse of a “value tell,” and the same “great-story-at-top-of-range” signature flagged on the richest names in the cohort.

Current enterprise value. ~112.5M shares × $118.86 ≈ $13.4B market cap. Adding the ~$7.5B of consolidated project debt and ~$0.4B parent debt, plus the $310M ABTC NCI, and netting ~$160M unrestricted cash and the ~$1.37B Bitcoin treasury (with most project-bond proceeds escrowed), the consolidated EV is on the order of ~$18–19B — though for a non-recourse, project-financed structure the cleaner lens is a sum-of-the-parts on the equity, valuing each campus net of its own ring-fenced debt.

EV per contracted MW — the cleanest cross-sectional anchor — shows HUT is the cohort’s most expensive.

Even crediting HUT for the best direct hyperscaler anchor and stripping out the ~$1.4B of ABTC-stake + Bitcoin value, HUT still screens at ~$28M/MW — the highest in the group. The market is paying a cohort-premium multiple and capitalizing a large slug of the uncontracted ~1 GW+ pipeline.

Sum-of-the-parts (illustrative; NO price target). Valuing the equity by parts:

• Contracted lease book: ~$1.1B of stabilized annual NOI (beginning ~mid-2027), capitalized at a fair stabilized data-center cap rate. At 7% (roughly the project-debt cost plus a thin spread) the gross asset value is ~$15.7B; at 6% (bull) ~$18.3B; at 8.5% (bear) ~$12.9B. Net of the ~$7.5B of project debt, the levered equity in the contracted book is roughly ~$5–8B, depending on the cap rate.
• ABTC stake: ~55% × ABTC’s (depressed) public market cap (~$0.9B) ≈ ~$0.5B, net of the NCI already on the balance sheet.
• Bitcoin treasury (HUT-attributable): ~$0.6–0.8B and shrinking.
• Legacy mining + Highrise AI + power pipeline option: a risk-adjusted ~$1–3B for the uncontracted ~1 GW+ pipeline, behind-the-meter optionality, and owned-GPU adjacency.
• Less parent net debt (small, near-zero after the Coatue convert force-converts).

Summing the contracted equity (~$5–8B) + ABTC (~$0.5B) + Bitcoin (~$0.7B) + a modest pipeline credit (~$1–2B) yields roughly $8–11B of equity value, or ~$70–95 per share, on the contracted book plus a modest pipeline credit — versus the ~$13.4B / $118.86 market price. The gap (~$25–50/share) is the aggressive value the market assigns to the unsigned pipeline, a benign Bitcoin tape, and an IG refinancing not yet achieved.

Scenario analysis (illustrative; NO price target; ~115–125M fully-diluted shares):

The current $118.86 sits above the base case, leaning toward the bull — i.e., the market is pre-paying for pipeline conversion and a successful IG refi and a cooperative crypto tape.

The implied-cap-rate crux. Capitalizing ~$1.1B of NOI against the whole-EV implies a cap rate in the high-5s-to-6% range — at or below the 6.1–6.2% cost of the very project debt funding the build. As with CIFR, a whole-EV implied cap at/below the debt cost means essentially all of the equity value rests on (a) the uncontracted pipeline converting and (b) an eventual IG refinancing that opens a positive levered spread to the common. Neither is yet in evidence.

A simple reverse-DCF frames how much pipeline is in the price. Take the contracted book at face value: ~$1.1B of stabilized annual NOI, capitalized at a 7% cap rate (a fair stabilized data-center multiple and roughly the cost of the project debt), is worth ~$15.7B of gross asset value. Net the ~$7.5B of non-recourse project debt and you have ~$8.2B of levered equity in the contracted book. Add the ABTC stake (~$0.5B net of NCI) and the shrinking Bitcoin stub (~$0.6–0.8B), subtract the small parent net debt (~$0 after the convert converts), and the signed, financed, in-construction business supports roughly $9–10B of equity, or ~$80–90 per share. To bridge from there to the ~$13.4B / $118.86 market price, the market must be assigning ~$3–4B of additional value — on the order of ~$25–35 per share, or a quarter-plus of the price — to the uncontracted ~1 GW+ pipeline, an IG refinancing not yet achieved, and Bitcoin upside. That can be the right call — power-constrained AI demand is real and HUT’s pipeline and anchor credit are among the best-positioned in the cohort — but it is unambiguously an option premium, not a margin of safety, and it is being paid at the 97.9th percentile of the stock’s own valuation history, mid-capital-cycle, by insiders who have never bought a share, on a 2.4-beta crypto security.

Embedded expectations — what must be true to justify $118.86: both leases must commence on schedule (~mid-2027) at the projected ~$1.1B NOI with tenants performing; the remaining ~650 MW of Beacon Point and the River Bend/FluidStack ROFO and the broader pipeline must convert to signed IG leases at premium rents; the 6.1–6.2% SPV notes must refinance toward investment grade; Bitcoin (and thus ABTC and the 2.4-beta equity) must not crater; and dilution must stay within the existing convert/ATM overhang. That is a demanding, multi-variable, multi-year set of conditions — achievable, but priced as if largely de-risked. This memo states no target and no recommendation; the embedded expectations are simply demanding.

11. Variant Perception

Consensus view. Sell-side is uniformly bullish — roughly 16 analysts, Strong Buy consensus, no Sells, with targets scattered by source and currency (some USD averages near spot, some stale pre-lease targets far below, some C$ targets implying upside). Short interest is meaningful at ~14–16% of the float — a sizable skeptic camp coexisting with the bullish sell-side, and a real squeeze dynamic. The Street narrative: a uniquely-positioned, vertically-integrated power-and-data-center platform that has landed the best direct hyperscaler anchor in the converted-miner cohort, ~$16.8B of contracted revenue, ~$7.5B of at-par institutional bonds, a ~1 GW+ pipeline, and optionality from ABTC and Bitcoin — “the year of execution.”

The strongest bull case. HUT pairs the highest-quality direct hyperscaler anchor in the cohort (a disclosed AA-or-higher tenant) with a Google-backstopped neocloud lease, on a multi-gigawatt power position that is hard to replicate quickly; the non-recourse, at-par project bonds prove institutional conviction; the parent is lightly levered (near-zero recourse debt after the convert converts) with ~$1.3B of liquidity; and there is real free optionality in the ~1 GW+ pipeline, the River Bend ROFO, Highrise AI, behind-the-meter gas, and a ~55% stake in a public Bitcoin accumulator. If the pipeline converts and the bonds refinance to IG, the equity is worth multiples of today, and Bitcoin upside is a call option on top.

The strongest bear case. The equity already pays the bull price. The ~$1.1B NOI is unaudited, not yet earning until 2027, and capitalized at/below the cost of the project debt; the stock is at the 97.9th percentile of its own history and the richest EV/MW in its cohort; ~$25–50/share of the price is unsigned pipeline; the company has never generated positive cash flow, diluted relentlessly at ~$18.61, and is run by management paid on Bitcoin-inclusive “Adjusted EBITDA” with a ~45%-against say-on-pay and zero insider buying; the ABTC subsidiary carries Trump-related governance/headline risk; and the whole thing is a 2.4-beta crypto security that lost ~95% in the last winter. A 2027 commencement slip, a failed IG refi, a neocloud-credit wobble, or a Bitcoin crash re-rates the equity hard.

The 3–5 assumptions that matter most (and what falsifies each):

1. Both leases commence on time (~mid-2027) at ~$1.1B NOI. → Falsified by a slipped Beacon Point/River Bend commencement or a tenant renegotiation in 2026–27.
2. The ~1 GW+ pipeline converts to signed IG leases. → Falsified by 2027 passing with no new signed lease beyond the current two (the remaining Beacon Point/River Bend capacity unleased).
3. The SPV bonds refinance toward IG, opening a levered equity spread. → Falsified by the next financing pricing wider, or rates/credit backing up.
4. Tenant credit holds (especially FluidStack/Anthropic). → Falsified by a neocloud-sector credit event or disclosure that the anchor is weaker than billed.
5. Bitcoin does not crater (ABTC + treasury + the 2.4-beta equity). → Falsified by a crypto-winter drawdown.

The factor-positioning read sharpens the variant perception. FactorsToday loads HUT overwhelmingly on a “Crypto Powerhouse Stocks” basket (beta 2.40) with a Market beta of ~2.2–2.4, clustering it with crypto ETFs (BITQ/DAPP/STCE) and HIVE/MARA/RIOT. Its recent risk-adjusted record is the signature of a crowded, high-beta, one-way-street-up momentum trade: a 1-year return of ~+547% (Sharpe ~5.3) and a violent multi-month tripling — but a 5-year maximum drawdown of −95%, the reminder that this exact security was nearly wiped out in the last crypto winter. That is evidence consensus is offsides not on business quality but on price and the durability of the move: the tape is pricing flawless execution and a benign crypto regime simultaneously. Where we land relative to consensus: the bull business case is largely correct and the execution is genuinely strong — but consensus is under-weighting the price, the capital-cycle timing, and the crypto-beta tail. The variant perception is “the equity is priced for a flawless, fully-converted gigawatt pipeline and an IG refi, at the 97.9th percentile of its own valuation history, on a 2.4-beta crypto security, mid-cycle.”

12. Fact vs. Interpretation Table

13. Open Questions

1. Who is the Beacon Point tenant? Confirmed AA-or-higher, multi-trillion-$ cap, not Anthropic/Google/FluidStack — but unnamed and unverifiable; its identity and standalone credit matter to the anchor quality.
2. Lease accounting: how will the triple-net leases be classified (operating vs. sales-type under ASC 842) once commencement hits, and what is the reported revenue/NOI shape?
3. Per-asset NOI economics (rent/kW/month, escalators beyond the 3% headline, term-end residual) — the ~$1.1B is a portfolio figure; per-lease spreads are undisclosed.
4. Will the SPV bonds refinance toward investment grade, and on what timeline? The entire levered-equity-spread thesis hinges on it.
5. Pipeline conversion: will the remaining ~650 MW of Beacon Point and the River Bend/FluidStack ROFO (+1,000 MW) sign IG tenants, and when?
6. Capitalized-interest treatment of the project bonds during construction, and pro-forma coverage once both projects energize.
7. Was the $200M Coinbase facility (June-2026 maturity) repaid, and did the Coatue convert force-convert as expected (parent-recourse debt ≈ zero)?
8. ABTC trajectory: further dilution of HUT’s ~55% stake, the Trump-related governance arrangements, and the political/headline risk path.
9. How much further ATM dilution (HUT and ABTC) is contemplated through the 2027 commencement gap?
10. Behind-the-meter generation and Highrise AI: concrete projects/financing, or still optionality?

14. What Must Be True (Bull and Bear)

Bull case — what must be true:

• Both signed leases commence on schedule (~mid-2027) at the projected ~$1.1B aggregate NOI, with tenants performing.
• The remaining Beacon Point/River Bend capacity and the broader ~1 GW+ pipeline convert to signed investment-grade leases over 2026–2029.
• The 6.1–6.2% SPV bonds refinance toward investment grade, opening a positive levered spread to the common.
• Bitcoin holds or rises (supporting ABTC, the treasury, and the 2.4-beta equity); dilution stays within the existing overhang.
• Falsification test: If, by year-end 2027, either Beacon Point/River Bend have not reached stabilized contracted NOI OR not a single new pipeline lease beyond the current two has been signed, the bull thesis is broken — the market will have paid a full-pipeline price for a stalled pipeline.

Bear case — what must be true:

• A 2027 commencement slip and/or the project-bond interest wall opens a funding gap that forces dilutive equity or a distressed refinancing.
• The pipeline stalls (rent compression from the capital flood; ERCOT/interconnection friction) and the cohort-richest ~$30M/MW proves unsupportable.
• A tenant-credit event (FluidStack/Anthropic neocloud stress) or a Bitcoin crash (ABTC + treasury + 2.4-beta equity) impairs value.
• Falsification test: If, through 2027, both leases commence on time at projected NOI, at least one major additional pipeline lease signs an IG tenant, the SPV debt refinances at/inside current spreads, AND Bitcoin holds, the bear thesis is broken — the equity will have de-risked into its valuation.

The two falsification tests are near-mirror images, which makes HUT an unusually datable security: the next 12–24 months of commencement dates, pipeline signings, refinancing spreads, and the Bitcoin tape will resolve the debate decisively.

15. Source Appendix

See the accompanying source appendix (HUT_source_appendix.md) and Appendix B of the combined report for the full, dated list of primary and secondary sources relied on in this memo, including: Hut 8 Corp. SEC filings (FY2024 & FY2025 10-K, FY2023 transition-period 10-KT, 10-Qs through Q1-2026, the 39-document 8-K corpus including the April-2026 River Bend and June-2026 Beacon Point bond filings, the DEF 14A proxy, the 66-document Form 4 insider corpus, and the S-4/425 American Bitcoin–Gryphon merger filings); the Q4-2025 and Q1-2026 earnings-call transcripts; ROIC.ai fundamental data; the AZI own-history valuation index and news feed; and the FactorsToday factor model. Every non-obvious fact in the memo carries an inline attribution; management commentary is treated as hypothesis and validated against filings and external data throughout.

APPENDIX A — Standard Diligence Questionnaire

Hut 8 Corp. (NASDAQ: HUT) — supplemental to the research memo. Report date: 2026-06-13. Answers are grounded in the underlying research; Fact / Interpretation / Assumption labels applied where it matters. Where a question does not map to the business model, the correct analog is given.

General

What thoughtful questions have other investors asked about this company?

• Is Hut 8 a Bitcoin proxy or a data-center landlord? (The whole re-rating thesis turns on the market accepting the latter while GAAP and the 2.4 factor beta still say the former.) (Interpretation.)
• Who is the unnamed AA-rated Beacon Point tenant, and is the disclosed rating real? (Open Question.)
• How much of the ~$119 price is the uncontracted pipeline vs. the signed book? (Our SOTP: ~$25–50/share is unsigned pipeline + IG-refi + Bitcoin optionality.) (Interpretation.)
• Will the 6.1–6.2% SPV bonds refinance toward investment grade, and is there a real levered equity spread? (Open Question.)
• What is HUT’s true ownership of ABTC after the ATM/PIPE dilution (≈55%, not the formation 80%)? (Fact.)
• Does the Trump-family ABTC affiliation create governance/headline/political risk? (Interpretation.)

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? GAAP “earnings” are meaningless — dominated by non-cash Bitcoin fair-value marks (a +$509M gain in FY2024 flipped to a −$220M loss in FY2025). The cash business has never been profitable on a free-cash basis. The contracted-NOI earnings engine (~$1.1B) has not started (begins ~mid-2027). So neither “high” nor “low” — pre-inflection. (Interpretation.)

Driven by external environment or internal actions? Both: the legacy result is driven externally (Bitcoin price, network difficulty, halving); the forward story is driven internally (signing leases, building, financing). (Interpretation.)

How stable are revenues? Today: volatile (Bitcoin-economics, recognized as coin-in-kind). Tomorrow: highly stable if delivered (15-year triple-net take-or-pay, IG counterparties, 3% escalators). The stability is prospective. (Fact/Interpretation.)

Outlook for products/services; how big is the market? The AI-data-center power market is large and growing (~$700B/yr hyperscaler capex), supply-constrained on power; the Bitcoin-mining market is commoditized and shrinking in relevance for HUT. Domestic (US/Canada) with the contracted book in Texas and Louisiana. (Fact/Interpretation.)

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More — capital is flooding the AI-data-center build (miners, neoclouds, REITs, PE, hyperscaler self-build); Marathon mid-to-late capital cycle. (Interpretation.)

How profitable is the business (ROIC, ROE)? Not yet meaningfully measurable on the HPC business (pre-revenue); GAAP ROE is noise (BTC marks, e.g. −189% FY2025). Stabilized triple-net NOI margins are projected >99% (tenant bears opex), but levered project equity returns depend on the cap-rate-vs-debt-cost spread and an IG refi. (Interpretation.)

How profitable is the industry — competitors, barriers to entry? Data-center landlording earns attractive stabilized economics today (cyclical-high), but barriers (power, interconnection, capital, speed) are replicable head-starts, not franchises. Bitcoin mining earns near-cost-of-power returns with no barriers. (Interpretation, Greenwald.)

Can the business be easily understood? Moderately — but the consolidated structure (data-center SPVs + a ~55%-owned listed Bitcoin miner + a Bitcoin treasury + an owned-GPU neocloud + power JVs) is messier than a pure landlord (CIFR) or pure treasury (MSTR). (Interpretation.)

Undermined by foreign low-cost labor? No — the asset is domestic power + interconnection + buildings; labor is a minor cost. (Fact.)

Do brands matter? Nature of competition? Switching costs? No consumer brand. Competition is for power and tenants. Switching costs are high post-signing (15-yr leases) but zero on the uncontracted pipeline. ABTC provides captive internal demand. (Interpretation.)

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The power-origination pipeline / interconnection rights and the development optionality (remaining ~650 MW Beacon Point, River Bend ROFO) carry little/no book value but are where the equity premium sits. (Interpretation.)

Off-balance-sheet liabilities? The project bonds are on the consolidated balance sheet but non-recourse to the parent (ring-fenced in SPVs). Construction completion obligations and the deep-ITM Coatue convert (~$300M unrecognized economic dilution) are the items to watch. (Fact.)

How conservative is the accounting? Mixed. Bitcoin is now fair-valued (ASU 2023-08, appropriate but volatile). The non-conservative tell is management’s “Adjusted EBITDA” including the Bitcoin mark and excluding growing SBC — flattering in up-markets. (Interpretation.)

How CapEx-hungry is the business? Extremely — gigawatt-scale data-center construction at ~$9–11M/MW, ~95% debt-funded; FCF deeply negative (~−$382M FY2025) and ~100% externally financed through 2027. (Fact.)

Capital Allocation & Management

How much FCF does the business generate; how is it used; philosophy? Negative FCF; the “philosophy” is to fund the build with non-recourse project bonds (the recent, favorable pivot) plus equity ATM (the legacy, dilutive habit), and to carve mining into the self-funding ABTC. (Fact/Interpretation.)

Significant acquisitions recently? The ABTC/Gryphon reverse merger (Sept 2025, brought mining public); the TransAlta divestiture of the Ontario gas plants (early 2026). (Fact.)

Buying back shares? No — and it would be inappropriate for a cash-burning developer. The company is a net issuer. (Fact.)

Issuing large amounts of new shares to insiders? SBC ballooned to $57.8M FY2025 (+178% YoY); serial ATM at a ~$18.61 average price; the Trump-affiliated founders hold ~45% of ABTC. (Fact.)

Compensation policy / motivations of management? LTI keyed to “Adjusted EBITDA including Bitcoin fair value” — a Bitcoin-linked metric, poorly aligned with the infrastructure-landlord narrative; ~45% say-on-pay opposition. CEO Genoot is the architect of the transformation; no insider has bought stock on the open market, and directors sold into the 2026 run. (Fact/Interpretation.)

Valuation & Market Data

ADR, MLP, or K-1 issuer? No — Hut 8 Corp. is a US-domestic C-corp (dual-listed Nasdaq/TSX), issues a 1099, not a K-1. (Fact.)

Dividend policy? None; no dividend (capital is consumed by the build). (Fact.)

How profitable is the business? See above — GAAP is Bitcoin-mark noise; cash profitability is negative; the durable profit engine (~$1.1B NOI) is prospective (2027). (Interpretation.)

Is net income diverging from cash from operations? Massively and in both directions — FY2024 showed +$332M net income on −$69M OCF; FY2025 showed −$226M net income on −$139M OCF. The divergence is the Bitcoin mark plus the non-cash coin-in-kind revenue. Always read the cash flow, not the P&L. (Fact.)

Risks & Downside

What factors would cause the stock to decline? A Bitcoin crash (2.4 beta; ABTC + treasury marks); a 2027 lease-commencement slip against the interest wall; a failed/wider IG refinancing; tenant-credit stress (FluidStack/Anthropic); pipeline non-conversion; ABTC/Trump headline risk; multiple compression from the 97.9th percentile; further dilution. (Interpretation.)

Risk of a catastrophic loss? Lower than it looks for a levered developer, because the project debt is non-recourse and ring-fenced in SPVs and the parent is lightly levered (≈zero recourse after the convert converts). But a severe equity drawdown (50%+) is entirely plausible on a Bitcoin-and-AI-narrative shock given the 2.4 beta and the rich entry. (Interpretation.)

Chance of a total loss? Low. Non-recourse structuring, parent liquidity (~$1.3B), and at-par institutional bonds make a zero unlikely; the realistic downside is a large re-rating, not a wipeout — though the −95% 5-year drawdown is a reminder of what crypto winters can do to this exact security. (Interpretation.)

Recent News & Events

Has the business environment changed recently? Yes, dramatically and favorably on signings/financing: two anchor leases (~$16.8B), ~$7.5B of at-par non-recourse bonds, a new Board Chair (E. Stanley O’Neal, June 2026), a new Head of IR. The stock jumped ~30–37% on the Beacon Point lease but fell on the June bond pricing. (Fact.)

Significant acquisitions / accounting changes / new markets, facilities, management? ABTC/Gryphon merger and consolidation; ASU 2023-08 Bitcoin fair-value adoption; new facilities (River Bend LA, Beacon Point TX) and the Ontario divestiture; CEO Genoot (2024) and a refreshed board/IR. (Fact.)

This questionnaire is supplemental and is not counted toward the memo’s length standard. It applies the Greenwald (Competition Demystified) and Marathon (Capital Returns) lenses where they add insight, consistent with the memo.

APPENDIX B — Source Appendix

Hut 8 Corp. (NASDAQ: HUT) — sources relied on in the research memo and diligence appendix. Report date: 2026-06-13. Primary sources prioritized over secondary; management commentary treated as hypothesis and validated against filings and external data. Access dates 2026-06-13 unless noted.

1. Company SEC filings (primary) — CIK 0001964789

Mirrored locally to output/HUT/sources/ (60-month corpus, 94 documents). Key documents:

• Form 10-K — FY2025 (filed early 2026): segment structure (Power / Digital Infrastructure / Compute), revenue $235.1M, the Bitcoin fair-value “loss (gain) on digital assets” reconciliation, Adjusted EBITDA reconciliation, Bitcoin treasury holdings/encumbrances, ABTC consolidation and noncontrolling interest, SBC, debt schedule, power-under-management (~1,020 MW / 15 sites).
• Form 10-K — FY2024: prior-year comparatives, the +$509.3M FY2024 Bitcoin gain, ~99.5M shares, ATM issuance history.
• Form 10-KT — FY2023 transition period (June→December fiscal-year change).
• Form 10-Q — Q1-2026 (and prior 10-Qs): Q1-2026 revenue $71.0M (+226% YoY), ~112.5M shares, balance sheet (cash $160M, Bitcoin ~15,679 coins ~$1.37B across custody/pledged lines, PP&E $812M, goodwill $209M, NCI $310M, net debt ~$245M pre-Beacon-bond), Q1-2026 SBC ~$50.9M, TransAlta gain ~$33.6M.
• Form 8-K — River Bend project bond (~April 2026): Hut 8 DC LLC, $3.25B 6.192% senior secured notes due 2042, non-recourse, ~95% LTC, ~$184M equity recovery; FluidStack lease with Google backstop.
• Form 8-K — Beacon Point project bond (~June 2026): Beacon Point DC LLC, $4.25B 6.129% senior secured notes due 2042, non-recourse, escrowed.
• Form 8-K — Beacon Point lease (~May 2026): 352 MW IT, 15-year, ~$9.8B base / ~$25.1B with renewals, 3% escalator, AA-or-higher unnamed hyperscaler, NVIDIA DSX architecture, Jacobs EPCM.
• Form 8-K — leadership/governance: E. Stanley O’Neal appointed Board Chair (June 11, 2026); Head of Investor Relations appointment (June 2026); CFO/officer changes.
• Form 8-K corpus (39 docs): earnings releases, the TransAlta Ontario divestiture, Bitcoin-debt refinancing (Coinbase→FalconX), DIFC license, ABTC matters.
• DEF 14A / DEFA14A proxy (3 docs): executive compensation — PSU plan keyed to “Adjusted EBITDA including Bitcoin fair value”; ~45% say-on-pay opposition; board composition.
• Form 3/4/5 insider corpus (66 Form 4s): no open-market purchases (code P); director sales into the 2026 run; CEO Genoot holdings (grants/merger stake, no buys/sells).
• S-4 / S-4/A / 425 (≈21 docs): American Bitcoin Corp formation and the Gryphon Digital Mining reverse merger (closed September 2025); ABTC ATM ($237.7M) and PIPE ($205.3M); Hut 8 ownership dilution from ~80% to ~55%.

EDGAR XBRL facts cross-checked via scripts/edgar.sh (revenue, net income, shares, debt, SBC, capex).

2. Earnings-call transcripts (primary management commentary; hypothesis)

• Q1-2026 earnings call (May 6, 2026) — via ROIC.ai: the $16.8B/~$1.1B-NOI two-lease framing; Beacon Point announcement and AA-rated unnamed tenant; River Bend/FluidStack/Google/Anthropic; the project-bond structure; ABTC as captive demand; Bitcoin-treasury wind-down; ~Q2–Q3 2027 commencement.
• Q4-2025 earnings call (February 25, 2026) — via ROIC.ai: FY2025 results, strategy, power portfolio, ABTC.
• Earnings-call enumeration: ROIC.ai list_earnings_calls (HUT, full history).

3. Quantitative data services (secondary; reconciled to filings)

• ROIC.ai — company profile, income statement, balance sheet, cash flow, profitability ratios, enterprise value, valuation multiples (FY2022–FY2025 + quarterly), latest stock price ($119.88 close 2026-06-12). Third-party aggregated; reconciled to EDGAR.
• AZI own-history valuation index (2026-06-12) — composite 97.9th percentile; P/E 76.2x (99.3rd), P/B 22.0x (94.9th), P/S 116x (99.6th); price $118.86. Own-history context only.
• AZI news feed — the June 5, 2026 $4.25B Senior Secured Notes pricing item; the Head of IR appointment item referencing $16.8B contracted lease revenue.
• FactorsToday factor model (2026-06-12) — stock loadings (dominant “Crypto Powerhouse Stocks” beta 2.40; Market beta 2.2–2.4; R² ~0.52); leaderboard (m3 / y1 / y3 / y5 risk-adjusted returns; y1 +547%, y5 max drawdown −95%); related-stocks cluster (BITQ/DAPP/STCE, HIVE, MARA, RIOT). Statistical estimates; interpretation labeled.

4. Peer / cross-read reports (comparable publicly-listed names, used for framing and comparison)

• CIFR (Cipher Digital) — 2026-06-13 (the master same-cohort cross-read: miner→landlord, EV/contracted-MW method, implied-cap-rate / reverse-DCF framing).
• WULF (TeraWulf) — 2026-06-12 (landlord-vs-neocloud distinction; FluidStack/Google/Anthropic credit cluster; warrant/derivative-mark QoE).
• IREN — 2026-06-10 (neocloud / owns GPUs).
• APLD (Applied Digital) — 2026-06-12 (build-to-suit landlord QoE; SPV/preferred-as-debt re-classification).
• MARA — 2026-06-13 and MSTR — 2026-06-10 (Bitcoin-treasury / mining comparables).

5. Public industry / secondary sources

• Hut 8 Corp. investor relations (hut8.com) — press releases on the leases, bonds, TransAlta sale, ABTC, governance.
• Trade and financial press (Benzinga, Yahoo Finance, etc.) — the June 5, 2026 bond-pricing and IR-appointment items; ABTC/Gryphon and Trump-family coverage.
• Public AI-infrastructure / data-center industry context (hyperscaler capex pace, ERCOT large-load interconnection, halving/hashprice dynamics) — used for the industry framing.

Every non-obvious fact in the memo carries an inline attribution. Where this memo and a filing disagreed on a material number, the filing governs; ROIC.ai/AZI/FactorsToday figures are third-party and were used as cross-checks, not as the primary record.