Galaxy Digital Inc. (NASDAQ: GLXY) — A Levered Crypto-Beta Book Re-Rated on a Single-Tenant Bet on CoreWeave’s Balance Sheet

Report date: 2026-07-11 Price (2026-07-10 close): ~$24.88 · Class A shares out: ~193M · Fully-exchanged units (incl. Class B/LP): ~389–393M · Economic market cap: ~$9.8B 52-week range: ~$16.84 – $42.86 · Beta (realized): ~2.8 · FY2025 GAAP net loss (to Class A): −$241M · Book value/share: ~$9.43 (P/B ~2.6x) Segments: Digital Assets · Data Centers (Helios) · Treasury & Corporate · CEO/Controller: Michael Novogratz (~49% voting power)

⚡ Claude’s Take

This block is the author’s own independent opinion and general information only — not investment advice, not a recommendation, and not a solicitation to buy or sell any security. The main body of this article (Sections 1–15) is written to take no position and sets no price target; the single exception is this block.

Call: HOLD / AVOID-here — and explicitly not-a-short. A genuinely valuable single asset (the Helios/CoreWeave lease) welded to a low-moat, mark-driven crypto-beta balance sheet, at a price that already capitalizes the good asset and gives you the volatile one for free-ish — with a controlling founder who sold $108M of stock into the top. Low-to-medium conviction. Directional zone: the crypto franchise + treasury are worth perhaps ~$4–5B of equity on their own (roughly where the whole company traded pre-re-rate in early 2025, ~$16–20/share); above that you are paying purely for Helios execution and CoreWeave’s creditworthiness. At ~$25 / ~$9.8B the base case is roughly fair, but the distribution is fat-tailed enough (~−70% / +80%) that I would not chase it here and would want a high-teens/low-$20s handle — where the CoreWeave optionality is closer to free — before the risk/reward turns attractive.

Two things are simultaneously true. (1) Novogratz made one genuinely excellent capital-allocation decision: he bought a stranded West Texas bitcoin mine for $65M in 2022, and turned it into Helios — an 800 MW gross / 526 MW critical-IT, 15-year lease to CoreWeave expected to throw off >$1B of average annual revenue (>$15B contracted) at REIT-like margins, ring-fenced behind a non-recourse-style $1.4B project facility. That single asset is the entire reason the stock re-rated from a ~$3.5B cap in Q1-2025 to ~$9.8B, and it is the reason this is not a short. (2) Everything wrapped around Helios is a low-moat, deeply cyclical crypto principal-dealer and sub-scale asset manager whose reported earnings are Bitcoin’s price wearing a financial-services costume: consolidated net income swung +$347M (FY24) → −$241M (FY25) entirely on the Treasury & Corporate mark-to-market book, while the operating franchise (Digital Assets) earns a real-but-tiny normalized ~$0.30–0.40 per fully-diluted share. You are buying two correlated bets — CoreWeave’s balance sheet and the crypto cycle — with no durable moat under either.

Framing: a momentum-cooled, very-high-beta (2.8) crypto-plus-AI-infrastructure proxy, not a data-center compounder. The factor model still clusters GLXY with the crypto complex (COIN, MARA, the BLOK/BITQ ETFs) even though the whole re-rate is a data-center story — the classic disconnect that is both the opportunity and the trap. The tape has cooled: ~42% off the $42.86 October-2025 all-time high, below all three key moving averages, 6-month relative strength negative. Bull trigger: a second, investment-grade, non-CoreWeave hyperscale tenant signs at Helios (converts single-tenant optionality into a diversified landlord) — I’d flip constructive. Bear trigger: a CoreWeave credit/AI-capex wobble, or a Bitcoin drawdown that impairs the Treasury book while the ~$1.2B/yr Helios build still needs external funding. Tag: “One great landlord bolted to a Bitcoin mood ring.”

📈 Stock Price Action — Five-Year Event Map

Over five years GLXY has run a full crypto-cycle round-trip and then some: from a ~$34 late-2021 bull-market peak, down ~92% to ~$2.50 in the FTX winter, back up more than 16x to an all-time high of $42.86 (Oct 21, 2025) on the Helios AI-data-center re-rate — then a ~61% drawdown to $16.84 (Feb 5, 2026) and a partial recovery to ~$24.88 today (2026-07-10). The stock now sits ~42% below its all-time high, mid-range within a wide 52-week band (~$16.84–$42.86), and just below its 21-, 50-, and 200-day moving averages (~$27.9 / $28.1 / $26.2) — a cooled momentum name. (Pre-2025 history reflects the TSX-Toronto listing back-adjusted; GLXY began trading on NASDAQ on May 16, 2025 after redomiciling — a currency/continuity gap applies to the earlier years.)

Cycle narrative. (1–2) GLXY is, at its core, a levered crypto beta — the 2021 peak and the ~90% 2022 collapse trace Bitcoin’s cycle amplified by the trading book’s FTX-era losses. (3) The 2023–24 recovery rode the spot-ETF approval and the post-election crypto surge, but the structurally important event was the 2024 signing of the 15-year CoreWeave lease at Helios — the seed of a new identity. (4–5) A tariff-driven April-2025 risk-off crash reset the stock to ~$8.70 just before the May NASDAQ re-listing broadened the shareholder base. (6) The August–October 2025 leg to the $42.86 all-time high was the AI-data-center re-rate proper: the $1.4B Helios financing, CoreWeave’s 800 MW commitment, and a $460M placement at $36 — notably including ~$135M of secondary stock sold by executives, including CEO Novogratz, days before the peak. (7) A ~61% drawdown into February 2026 unwound the momentum as crypto and AI-infra sentiment cooled together (crypto market cap fell ~20% in the quarter). (8) Helios Phase I’s on-schedule delivery and the Q2-2026 start of CoreWeave rent supported a partial rebound to the mid-$20s. The price moves are facts; the attributed drivers are interpretation. (No price target, no recommendation.)

1. Executive Summary

Galaxy Digital Inc. is a ~$9.8B-market-cap (economic) diversified digital-asset financial-services firm founded and controlled by Michael Novogratz, which in May 2025 redomiciled from Toronto (TSX) to Delaware and listed on NASDAQ. It reports three segments: Digital Assets (institutional crypto trading, OTC/derivatives, lending, structured products, capital-markets/M&A advisory, asset management, and the GK8 custody/tokenization stack), Data Centers (the Helios campus in West Texas, leased to CoreWeave for AI/HPC), and Treasury & Corporate (a proprietary Bitcoin treasury, a ventures/private-equity book, and legacy bitcoin mining).

The single most important fact about GLXY is that its entire re-rating over the past eighteen months — from a ~$3.5B market cap in early 2025 to ~$9.8B today — is the market capitalizing one asset: the Helios/CoreWeave lease. Galaxy bought a stranded bitcoin mine from Argo Blockchain for $65M in December 2022 and converted it into an 800 MW gross / 526 MW critical-IT AI data-center campus leased to CoreWeave under 15-year agreements (plus two five-year extensions) expected to generate >$1B in average annual revenue (>$15B over the term). Phase I (133 MW critical IT) was delivered on schedule with rent commencing in Q2 2026. The build is ring-fenced behind a $1.4B senior secured project facility (Galaxy Helios I LLC, bankruptcy-remote) plus a $1.3B convertible — disciplined structuring that insulates the parent.

That is the bull case, and it is real. The bear case is that everything else is a low-moat, deeply cyclical, mark-to-market crypto business. GAAP is close to uninterpretable: FY2025 “revenue” of $60.4B is ~99% gross principal-trading notional, offset by near-equal cost, producing a negative reported gross profit; the counterparty trading spread is only ~$61M. Consolidated net income swung from +$347M (FY2024) to −$241M (FY2025) almost entirely on the Treasury & Corporate mark-to-market book (segment pretax +$290M → −$463M), not on operations. The genuine operating franchise — Digital Assets — is improving (segment pretax −$54M → +$47M → +$194M across FY2023–25) but earns a normalized ~$0.30–0.40 per fully-diluted share, trivial against a ~$25 stock, and its industry (institutional crypto dealing + sub-scale asset management) is structurally poor and moat-light — a lower-quality “COIN-lite” dealer.

Capital allocation is a study in contrasts: the Helios pivot and the cheap GK8 distressed acquisition were excellent; the surrounding sprawl across five business lines, funded by relentless equity and note issuance (June-2025 ~$600M offering, October-2025 $460M placement, $1.3B convertible, a fresh shelf), is empire-building. Most tellingly, in the October placement at $36, Novogratz personally sold 3.0M shares for ~$108M and the President/CIO sold another ~$27M — days before the all-time high, with the stock now ~31% below that exit. Only one director bought open-market.

Valuation must be done sum-of-the-parts. On a sober base case, Helios (~$9.4B gross at ~12x stabilized NOI), the Digital-Assets franchise (~$3.0B), and the net treasury (~$1.8B), less ~$4.6B debt, imply ~$9.6B of equity — roughly the current cap. The stock is therefore approximately fairly valued on the base case, with a genuinely fat-tailed distribution driven by two correlated bets: Helios/CoreWeave execution-and-credit, and the crypto cycle. This memo takes no position; the embedded-expectations analysis in Section 10 is the crux.

2. Business Overview

Galaxy is two fundamentally different businesses wearing one ticker, and the GAAP statements obscure both. It is best understood as (a) a crypto principal-dealer and asset manager, (b) a single-tenant AI-data-center landlord, and © a proprietary crypto/ventures treasury — bundled under one Up-C holding structure.

2.1 The revenue mirage

Reported “Total revenues” — $60.4B (FY2025), $42.6B (FY2024), $51.6B (FY2023) — is a gross-flow artifact. Galaxy acts as principal in digital-asset trades, so the full notional of each sale ($59.85B in FY2025) is grossed up into revenue and ~fully offset by transaction cost ($60.2B), leaving a negative GAAP gross profit (−$116M in FY2025). The economically meaningful counterparty trading spread is only ~$61M. Any use of consolidated P/S or gross margin is therefore meaningless, and management itself steers on adjusted gross profit by line. The genuinely recurring, annuity-shaped revenue lines are small: Fees $158.6M (investment banking + asset management), gross interest income $147.9M (against $158.0M of borrowing cost — a thin net spread), net staking rewards ~$56M (gross $220M less $179M of distributions; Galaxy keeps only 5–10%), and mining ~$6.4M.

2.2 The three segments

Digital Assets (the operating core). Internally split into two units. Global Markets — institutional OTC spot and derivatives trading, lending, structured products, prime brokerage, plus investment-banking and M&A advisory — is the flagship, transacting with >1,600 counterparties across ~$12.3B of assets on platform (12/31/25). Asset Management & Infrastructure Solutions houses Galaxy Asset Management (funds and ETFs, ~$5B AUM), plus the “Infrastructure” stack of staking, tokenization, and GK8 institutional-grade custody technology. In Q1-2026, Digital Assets contributed ~$49M of adjusted gross profit (Global Markets ~$31M, +3% QoQ despite industry trading volumes down 25%+; Asset Management ~$18M). GalaxyOne, a retail fintech platform, launched in October 2025 — an additional (unproven) arena.

Data Centers (Helios — the re-rate engine, but ~$0 of revenue in FY2025 and Q1-2026). The Helios campus (Dickens County, West Texas panhandle; ~2,200 acres) was acquired from Argo Blockchain for $65M in December 2022 as a stranded bitcoin mine and converted to AI/HPC. CoreWeave has committed 526 MW of critical IT load (of 800 MW gross approved) across Phases I–III under 15-year leases with two five-year extension options, expected to generate >$1B in average annual revenue. Phase I (133 MW critical IT / ~200 MW gross) was delivered on schedule, with rent commencing Q2 2026; Phase II data halls arrive in 2027. ERCOT approved an additional 830 MW in January 2026, expanding approved capacity toward 1.63 GW (with a longer-run path management frames toward ~3.6 GW).

Treasury & Corporate (the crypto-beta book). A proprietary portfolio: a Bitcoin treasury, a ventures/PE book (Ripple, Forward Industries, Xapo, CPO, spot BTC/ETH ETFs ~$536M, own-fund LP interests ~$259M), and the run-off bitcoin-mining operation. This segment is the tail that wags the GAAP dog: its pretax result was +$304M (FY23), +$290M (FY24), −$463M (FY25), and it single-handedly turned FY2025 into a consolidated loss.

2.3 How it makes money — and how much is recurring

Mechanically: Global Markets earns bid-ask spreads and financing net-interest; Asset Management earns fees on AUM; Infrastructure earns staking/custody take-rates and licensing; the treasury earns (or loses) mark-to-market on crypto and ventures; and Helios will earn contractual rent. Only the asset-management fees, staking/GK8 take, and (prospectively) the Helios lease are annuity-shaped. Trading gains embed proprietary risk; advisory fees recognize point-in-time at deal close (lumpy); treasury results are pure crypto beta. This is a predominantly transactional and mark-driven revenue base with a thin recurring core — not a subscription or toll-road business.

Verdict: a coherent narrative (bridging TradFi and digital assets, now with an AI-infrastructure leg) sitting atop an incoherent earnings base — one small improving operating franchise, one pre-revenue single-tenant landlord, and one volatile proprietary crypto book. The reader must value the parts; the whole tells you nothing.

3. Industry Dynamics

GLXY straddles two industries with opposite structural characters, and is migrating capital toward the (currently) better one.

3.1 Crypto financial services / trading & asset management — structurally bad

Institutional crypto dealing is a low-barrier, price-taking, spread-compressing market. There is no meaningful barrier to entry: capital and technology are widely available; counterparties multi-home across dealers (Galaxy, Cumberland/DRW, B2C2, Wintermute, FalconX, Coinbase, Binance, Kraken); and there are no switching costs at the trade level. Spreads compress as the asset class institutionalizes. Cross-referencing Coinbase’s public disclosures, even the category leader shows a melting retail take-rate, five-year peak-to-peak revenue that is roughly flat, and ~56% pure crypto-beta earnings that cannot be underwritten quarter-to-quarter. Galaxy is the institutional/OTC analog with less moat than Coinbase — no regulated-exchange brand annuity, no large stablecoin-reserve float income, and a sub-scale ~$5B asset-management book against BlackRock’s IBIT, Fidelity, Grayscale and Bitwise, whose spot-ETF complexes run into the hundreds of billions.

The 2024–25 regulatory thaw (spot-ETF approvals, GENIUS/CLARITY-type legislation, a friendlier SEC) is a genuine demand tailwind — but it is a moat dilutor, not a moat builder: clearer rules pull TradFi incumbents (banks, CME/ICE, BlackRock) directly into Galaxy’s lanes. In Marathon’s capital-cycle terms, this is a high-return, capital-attracting phase that invites entrants and mean-reverts returns.

3.2 AI/HPC data-center development — good today, but a capital-cycle top

The Helios leg sits in the opposite structure. Demand for GPU-ready, high-density power runs far ahead of supply; the binding constraint is electrical power and time-to-energization (ERCOT interconnection, contiguous megawatts), not capital. An operator that already controls energized or fast-to-energize power — as Galaxy does at Helios via its ERCOT interconnect and behind-the-meter optionality — holds something genuinely scarce, and the rents are real.

But the AI data-center boom carries the unmistakable signature of a capital cycle at or near its peak: record industry capex, a stampede of entrants (hyperscalers building their own, established REITs Digital Realty and Equinix, Blackstone/QTS, Vantage, Switch, and a wave of crypto-miner converts — Core Scientific, IREN, TeraWulf, Cipher, Applied Digital, Hut 8), and abundant, increasingly aggressive financing. Marathon’s lesson is that supply, not demand, sets returns — and supply is being added furiously. The risk is not that AI demand is fake; it is that the duration and credit quality of the demand may not match the 12–15-year, highly-levered assets being built against it, and that a single air-pocket in hyperscaler capex re-rates the entire cohort. public analysis of Core Scientific develops this at length for the identical CoreWeave-landlord model.

Verdict: Galaxy earns its living in a structurally poor industry (crypto dealing/asset management) and has bought a call option on a structurally good-right-now one (AI data centers). “Good right now” in a furious capital cycle is not durably good. Investors should price the data-center leg as a cyclical at an up-cycle with single-tenant credit risk, not a secular compounder — and the crypto leg as a moat-light cyclical.

4. Competitive Position

Does Galaxy have a durable competitive advantage? In the core, no. At Helios, what exists is contractual backlog, not a franchise moat. Run the Greenwald taxonomy segment by segment.

4.1 Digital Assets — no moat

• Supply/cost advantage: none. Trading, OTC, lending, and structured products are commodity dealer activities; Galaxy has no structural funding-cost or technology edge over Cumberland, B2C2, Wintermute, FalconX or the exchanges.
• Demand-side captivity / switching costs: none. The >1,600 counterparties multi-home; a client can route the next block trade elsewhere at zero cost.
• Economies of scale + captivity (Greenwald’s strongest combination): absent. At ~$5B AUM in asset management, Galaxy is sub-scale against $100B+ ETF complexes; in trading it is a price-taker in a national/global market with many credible bidders.
• Intangibles / brand: the one asset with mild value is the GK8 custody technology and licensing, plus Novogratz’s brand and relationships (deal flow, capital-raising reach). Apply the tie-to-a-financial-outcome test: if the “moat” is one man’s rolodex, it does not survive his departure — it is not a moat.

4.2 Helios — a single-tenant call option, not a moat

There is real switching-cost captivity within the signed 15-year CoreWeave lease: once GPUs are installed, moving them is costly and disruptive, so the contracted cash flows are sticky. But that protects the signed revenue, not Galaxy’s ability to win the next tenant — which is the entire growth thesis beyond CoreWeave. Data-center revenue is ~100% CoreWeave, and CoreWeave is itself an unprofitable, heavily-levered GPU-neocloud, so the load-bearing risk is counterparty credit, identical to the Core Scientific setup (where analysts flagged ~67% CoreWeave concentration). A useful framing of the identical Core Scientific setup applies verbatim: “a moat that does not survive the next tenant negotiation is a backlog, not a moat.” Galaxy’s Helios is arguably better CoreWeave collateral than CORZ’s — a longer 15-year term, a positive-equity parent, and a ring-fenced facility — but it shares the same single-customer, single-neocloud-credit vulnerability.

4.3 Direct comparison

Versus Coinbase, Galaxy’s crypto franchise is lower-quality (a dealer, not a regulated exchange with a fee annuity and stablecoin float). Versus Core Scientific / IREN / TeraWulf, Helios is comparable-to-better on contract quality but shares the concentration risk and competes for the next tenant against hyperscalers and REITs with lower cost of capital than a crypto-adjacent developer. Versus Digital Realty / Equinix, Helios is faster and cheaper per MW but vastly weaker on customer diversification, balance sheet, and cost of capital.

Verdict: GLXY is a crowded-market, weakly-differentiated crypto dealer/asset manager (no durable advantage) that has re-rated on the optionality of one 15-year, single-tenant AI data-center lease — contracted cash-flow visibility, not a competitive moat. If you own GLXY you are underwriting a spread-compressing crypto-beta engine and CoreWeave’s balance sheet — neither of which is a franchise.

5. Growth History and Forward Opportunities

History — low-quality and impossible to read on the GAAP top line. Reported “revenue” ($51.6B → $42.6B → $60.4B, FY23–25) is gross trading notional and tells you only that principal-trading volumes fluctuate with crypto activity, not that the business is growing. The honest growth record is in the segment pretax bridge:

Two facts fall out. First, the operating franchise (Digital Assets) is genuinely improving — −$54M → +$47M → +$194M — a real inflection driven by trading spreads, derivatives, fees, lending and staking as the platform scaled counterparties. Second, the entire consolidated result is set by the Treasury & Corporate mark-to-market book, which is crypto-price beta, not a growth engine. So the multi-year “earnings growth” is largely a Bitcoin chart.

Forward opportunities — high potential, mostly unrealized.

• Helios ramp (the one that matters): from ~$0 revenue in FY2025 to >$1B average annual run-rate as Phases I–III energize (Phase I rent from Q2-2026; Phase II 2027). This is contracted, high-margin, and the highest-quality growth Galaxy has — but concentrated in one tenant.
• Uncontracted Helios megawatts: ERCOT-approved capacity toward 1.63 GW (vs. 800 MW contracted) is merchant-development optionality — real, but zero signed non-CoreWeave revenue as of mid-2026.
• Asset management / ETFs: scaling a sub-scale ~$5B book in a market dominated by trillion-dollar managers — plausible but not differentiated.
• GalaxyOne retail fintech: a new, unproven consumer arena launched Oct-2025 — optionality with execution and customer-acquisition risk.
• Tokenization / GK8 custody: a credible institutional-infrastructure play riding the RWA/tokenization theme, but small today.

Verdict: the quality of forward growth is bifurcated — the Helios ramp is high-quality but single-tenant and finite-life; the rest (asset management, GalaxyOne, tokenization) is speculative optionality in competitive markets. Underlying it all, the crypto-cyclical Treasury book will continue to dominate reported results in both directions. This is not a clean secular growth story; it is a contracted data-center ramp stapled to a crypto-beta balance sheet.

6. Financial Quality

One-sentence read: strip the marks and Galaxy is a small-but-improving operating franchise stapled to a levered crypto-beta balance sheet and a pre-revenue, capital-devouring data-center build — not yet a self-funding, cash-generative business.

6.1 Normalized earning power

The recurring, annuity-shaped income lines are modest: Fees $158.6M, net interest (a thin spread on $147.9M gross income against $158.0M borrowing cost), net staking ~$56M, and mining ~$6.4M. Even within Digital Assets, ~$864M of the FY25 segment result is “gains from operations” (trading/derivatives/digital-asset marks), so the pure fee+interest+staking annuity is far smaller than the segment profit implies. A rough mid-cycle normalized figure — Digital Assets operating income net of corporate overhead, Treasury marks zeroed, taxed normally — is only ~$100–150M to all unitholders, i.e. ~$0.30–0.40 per fully-diluted share against a ~$25 stock. The price is an option on Helios plus crypto beta, not a multiple of current cash economics.

6.2 Quality-of-earnings red flags

1. Net income = non-operating marks. The FY24→FY25 swing (+$347M → −$241M) is almost entirely the Treasury & Corporate book. Not a run-rate — in either direction.
2. Impairment/gain gross-up. FY2025 shows a $753.7M digital-asset impairment against a $778.2M “gain on digital intangibles” (gain on derecognition of impaired assets) — a large, two-way, non-economic artifact of the lower-of-cost accounting model for crypto that does not qualify for fair-value treatment. The net digital-asset economic result was roughly flat (~−$11M); both gross lines are noise.
3. Operating cash flow is negative every year (−$317M, −$205M, −$4M FY23–25) despite positive GAAP net income in FY23/FY24. Heavily caveated (trading-inventory and collateral movements flow through operating cash for a dealer), but the firm plainly does not self-fund: FY2025 free cash flow was −$1.5B (capex $1.19B = Helios), covered by +$2.4B of financing (~$851M net equity + ~$2.1B notes).
4. Serial dilution. June-2025 offering (~26.4M primary shares), October-2025 placement (9.0M new + 3.75M insider secondary @ $36), and $1.3B 2031 exchangeable notes (up to ~74.2M Class A on exchange). Diluted weighted shares were ~366M. A $200M buyback was authorized in February-2026 while raising equity and debt — pro-cyclical and cosmetic.
5. Up-C / TRA / offshore tax. GDI owns <50% of the operating partnership (GDH LP); a Tax Receivable Agreement pays insiders (incl. Novogratz) 85% of realized tax savings. Domestic pretax income is persistently negative (−$359M FY25) while foreign is positive (+$89M); economic income sits offshore and is largely untaxed (FY25 tax benefit of −$29.3M). ROE/ROIC are near-meaningless quality signals here — reported ROE (−19% / +24% / +50% across FY23–25) is just leveraged crypto beta.
6. Concentration. Two trading platforms = ~38% of revenue; three counterparties = ~45% of the loan book; one platform custodies ~14% of digital assets.

6.3 Balance sheet

Total assets $11.35B; total equity $3.03B, of which NCI (Class B/LP) is $1.08B, leaving equity attributable to Class A of $1.956B. Intangibles + goodwill are only ~$93M, so tangible ≈ book: roughly $10.26 per Class A share (equity-before-NCI ÷ ~190.7M) or ~$7.80 fully-exchanged (total equity ÷ ~389M units) — bracketing the ~$9.43 quoted book value; P/B ~2.4–3.2x. Digital assets of $5.83B (assets) are largely matched by $4.38B (liabilities); net directional Treasury crypto is only ~±$184M per ±20% move (a ~$0.9B net long). Property & equipment jumped to $1.42B (from $237M) on the Helios build.

Debt is best read in two buckets. Recourse corporate debt ~$3.03B: exchangeable notes ~$2.15B principal (2026 $500M @ 3.0% / ~7.0% effective; 2029 $402.5M @ 2.5% / ~9.2%; 2031 $1.3B @ 0.5% / ~0.9%) plus an ~$878M drawn balance of the $1.4B Deutsche Bank senior secured term loan financing Helios. On top sits ~$4.45B of callable, collateralized trading funding (margin loans + collateral payable) — leverage against a matched trading book, not corporate recourse. Cash is $1.25B; near-term calls include ~$530M of data-center commitments and a ~$160M legal-settlement tail. Management concedes it will need additional debt/equity to complete the Helios conversion.

Verdict: economics do not “improve with scale” in any conventional sense, because reported profitability is not operating profitability — it is crypto-price beta booked through a proprietary balance sheet. There is a real, improving trading/asset-management franchise underneath, but its normalized cash earning power (~$0.30–0.40/share) is trivial against the price, and the data-center segment — the entire re-rating narrative — is still ~$0 of revenue and a net consumer of ~$1.2B/yr of capital. This is a levered, mark-driven crypto-beta balance sheet with a genuine but small fee/trading annuity, plus a pre-revenue data-center call option, funded by continuous issuance rather than internal cash flow.

7. Capital Allocation

Verdict up front: one brilliant decision surrounded by dilution-funded empire-building, run by a controller who has been a heavy net seller.

7.1 The good

The Helios pivot is a genuinely excellent capital-allocation move: a stranded bitcoin mine carried near $65M was converted into a 15-year, 526 MW CoreWeave lease with >$1B average annual revenue (>$15B contracted). Crucially it is project-financed and ring-fenced — the bankruptcy-remote Galaxy Helios I LLC drew a $1.4B Deutsche Bank senior secured term loan (Aug-2025), supplemented by a $1.3B low-coupon convertible (Oct-2025, exchange price ~$55.76) and October equity — so the AI-infrastructure bet is insulated behind non-recourse-style debt rather than funded off the parent. The GK8 custody acquisition (~$44M, Feb-2023, out of the Celsius bankruptcy estate at ~60% below Celsius’s $115M cost) was likewise a cheap, coherent distressed buy.

7.2 The rest — empire-building funded by issuance

Everything else looks like sprawl across five arenas — crypto OTC/derivatives dealer, asset management, AI data centers, BTC treasury + ventures/PE, and retail fintech (GalaxyOne) — plus advisory mandates like the K Wave Media crypto-treasury partnership (Jul-2025; Galaxy as asset manager/advisor and a small stakeholder, giving KWM access to ~$1B of institutional capital — not a $1B Galaxy outlay). The core dealer franchise is low-moat, mark-driven and cyclical; the headline Q3-2025 $505M net profit was crypto marks plus a +140% trading-volume spike, not recurring earnings. Buybacks exist on paper (the $200M February-2026 authorization) but are immaterial and pro-cyclical — GLXY is a massive net issuer, not a returner of capital.

7.3 Dilution & the Up-C control map

• Structure: an Up-C with one vote per share (not super-voting). Class A trades on NASDAQ/TSX; Class B (held by Novogratz-controlled Galaxy Group Investments) carries the vote but no economics, tracking LP Units in GDH LP that are redeemable 1:1 into Class A.
• Control: Novogratz holds ~192.4M shares = ~50.1% of as-converted Class A and ~49.25% of combined voting power; the board is effectively his (all six nominees elected May-2026). The as-converted share base is ~389–393M, cross-checking to the ~$9.8B economic cap at ~$24.88.
• 2025 dilution stack: May redomicile/US listing → June offering (~35.98M shares, ~$600M gross) → October placement (~9.0M primary @ $36, ~$325M net) → $1.3B convertible (dilutive above ~$55.76) → a May-2026 S-3ASR shelf for further capacity. SBC is heavy (Feb-2026 grants: Novogratz 174k, others in the ~97–174k range).

7.4 Insider behavior — decisively net sellers (a bearish tell)

The critical fact: the October-2025 “$460M strategic investment” at $36 was ~$325M of primary capital to the company plus ~$135M of cash-out to insiders — Novogratz sold 3.0M shares for ~$108M and President/CIO Chris Ferraro sold ~$27M, days before the $42.86 all-time high, with the stock now ~31% below that $36 exit. Across the corpus, only one insider bought open-market: Director Douglas Deason (~$1.55M) — no officer, and not Novogratz, bought a single share, and apart from Deason none re-entered on the pullback. All other named-officer dispositions are routine RSU tax-withholding or cashless option exercises. Net read: a strongly bearish insider signal layered on top of the controller’s $108M top-tick sale.

7.5 Compensation

Novogratz’s 2025 total compensation was ~$7.65M (salary $500k reinstated after taking $0 in 2019–23; a ~$2.92M discretionary cash bonus; ~$4.18M equity). Bonuses are discretionary with no disclosed formulaic ROIC/EPS grid — alignment rests on his ~$4.8B stake and stock-ownership guidelines, not on pay design, and that alignment is undercut by the $108M sale.

Verdict: Helios is genuinely value-additive and well-financed; GK8 was smart and cheap. But the surrounding deployment is scattershot, funded by relentless equity and note issuance while the controlling founder cashed out $108M near the peak — a record best described as one great bet inside a dilution machine run by a heavy seller.

8. Changes and Headwinds — Last Two Years

Strategic / structural:

• May-2025 redomicile and NASDAQ listing. Moved from TSX Toronto to a Delaware holdco with an Up-C structure; broadened the shareholder base and enabled US index/ETF inclusion — a genuine liquidity and cost-of-capital event.
• The Helios/CoreWeave build-out. The 15-year lease, CoreWeave’s escalation to 526 MW committed (Apr & Aug 2025), the $1.4B project facility (Aug-2025), the ERCOT +830 MW approval (Jan-2026), Phase I delivery, and Q2-2026 rent commencement — the defining transformation of the company’s identity and valuation.
• GalaxyOne retail launch (Oct-2025) and the K Wave Media advisory mandate (Jul-2025) — new arenas.

Capital markets: June-2025 ~$600M offering; October-2025 $460M placement at $36 (with ~$135M insider secondary); $1.3B convertible (Oct-2025); a May-2026 shelf. Continuous issuance is itself a headwind to per-share value.

Regulatory / litigation: the New York Attorney General’s LUNA/Terra settlement (reached ~March-2025, ~$200M reported) is a legal one-timer tied to Galaxy’s 2021 promotion of Terra — a reputational and cash overhang now largely resolved. The broader 2024–25 regulatory thaw is a demand tailwind but a competitive dilutor (Section 3).

Leadership/board: the annual meeting (May-2026) re-elected all six nominees; the structure keeps Novogratz firmly in control.

Verdict: the changes are net thesis-strengthening on the asset side (Helios is real and executing) but thesis-weakening on the per-share side (relentless dilution, insider selling, and a still-cyclical crypto core). The transformation is real; the shareholder-value translation is diluted and controller-dominated.

9. Risk Analysis

Catastrophic-loss assessment: a total loss is unlikely — the Helios debt is ring-fenced, the trading book is largely matched/collateralized, and net directional crypto is modest (~$0.9B net long). But a severe (−50%+) drawdown is a live scenario if (a) CoreWeave’s credit wobbles, (b) Bitcoin enters a deep winter while the Helios build still needs external funding, or © the AI-infra cohort re-rates — and these risks are correlated (a risk-off event hits crypto, AI-infra sentiment, and financing access simultaneously).

10. Valuation Discussion (Embedded Expectations)

GLXY cannot be valued on consolidated multiples: GAAP EPS is a loss, “revenue” is gross notional (EV/sales ~0.12x is meaningless), and net income is dominated by non-cash marks. The only defensible frame is sum-of-the-parts, because the three segments are economically unrelated.

The anchor — Helios. CoreWeave has committed 800 MW gross / 526 MW critical IT across Phases I–III under 15-year leases, expected to generate >$1B average annual revenue at REIT-like margins. This is the identical business model seen at Core Scientific (CORZ): a single-tenant, long-dated, depreciating colocation asset leased to the same sub-IG, cash-burning neocloud. The Core Scientific comparison sets the multiple discipline — a contracted-but-single-tenant, finite-life stream deserves ~12–16x stabilized unlevered operating profit, a clear discount to a diversified IG data-center REIT’s ~20–25x. On ~$800M stabilized NOI (>$1B rent × ~80% margin), that supports ~$9–13B of gross asset value at full ramp — but only Phase I is energized today, and Phases II–III need several billion more of capex.

Digital-Assets franchise — valued on normalized cross-cycle fee/spread earnings, not a peak year (a COIN-lite dealer). Treasury — the owned BTC + ventures + PE, net of related liabilities, marked to market (a fraction of the gross $5.8B “digital assets,” most of which is a matched trading book).

Sum-of-the-parts (gross asset value, then net total debt ~$4.6B; vs. the ~$9.8B economic cap):

Embedded expectations. At ~$9.8B the market is pricing the base case: Helios fully executes at a mid-single-tenant ~12x multiple, the crypto franchise normalizes, and the treasury holds. Roughly half the base-case gross value (~$9.4B of $14.2B) is Helios — the crypto franchise plus treasury alone (~$4.8B gross) approximates GLXY’s entire pre-US-listing valuation (~$3.5B cap in Q1-2025). In other words, the whole re-rate from ~$3.5B to ~$9.8B is the CoreWeave AI-data-center story being capitalized. The market is underwriting three things: (1) all Helios phases get built and energized on time; (2) CoreWeave pays for 15 years; and (3) the crypto book survives drawdowns without impairing capital. What looks under-discounted is the single-tenant CoreWeave credit concentration and the generous implied Helios multiple relative to the discount the identical CORZ asset carries.

On its own history the stock is rich: valuation-percentile analysis puts P/B at the 77th percentile (2.64x), P/S at the 85th, composite 81st of its own multi-year range (P/E null on GAAP losses). But “rich vs. own history” largely reflects that the market now capitalizes Helios’s contracted future earnings well above its depreciated carrying cost — book value doesn’t capture a 15-year lease. The most bullish external mark (a sell-side SOTP near ~$19.4B, “48% discount”) is essentially our bull bookend — it stacks a rich Helios and a rich franchise simultaneously — and should not be read as fair value.

Verdict: at ~$9.8B the equity is roughly fairly valued on a sober base case, with a genuinely fat-tailed distribution (~−70% / +78%) driven by two correlated bets — Helios/CoreWeave execution-and-credit, and the crypto cycle. No price target.

11. Variant Perception

Consensus view. The bullish street narrative is that GLXY is a mispriced AI-infrastructure play: a REIT-quality, 15-year, >$1B/yr contracted cash-flow stream (Helios) buried inside a stock the market still trades as leveraged crypto beta — so the sum-of-the-parts is worth far more than the ~$9.8B cap (the ~$19.4B sell-side SOTP). The bearish counter-narrative is that it is a low-moat crypto dealer with a volatile balance sheet, a single-tenant data-center bet on a shaky neocloud, relentless dilution, and a founder who sold $108M into the top.

The factor-positioning read sharpens the debate. On the factor model GLXY is unambiguously a very-high-beta crypto-plus-AI-infra proxy, not a data-center REIT: realized beta ~2.83, strongly negative LowVolatility (−1.55) and Liquidity (−0.95) loadings, with returns explained by Market (+2.12), Fintech (+1.67), a “Crypto Powerhouse” basket (+0.84), Broker-Dealers (+0.70) and CreditRisk (+0.39) — while the classic style factors (Momentum, Value, Quality, Growth) are all zeroed out (r² ~0.48; ~52% idiosyncratic). Its factor-nearest neighbors are crypto ETFs (BLOK 0.95, BITQ 0.91, BKCH 0.91, GBTC 0.89) and single-name crypto/miner betas (COIN 0.88, MARA 0.88, HUT 0.87, IREN). The tape still clusters GLXY with the crypto complex even though the entire re-rate is an AI-data-center story — the core disconnect and the crux of the variant perception. Momentum has cooled but not broken: 12-month relative strength still positive (+23.4) with positive alpha (+0.17), but the stock is ~42% off its RS peak, 6-month RS has turned negative (−2.5), and it trades below all three key moving averages. Neither a clean one-way-street-up momentum trade nor yet a falling knife — a decelerating high-beta proxy whose next major move is dictated by Bitcoin’s direction and Helios/CoreWeave execution in roughly equal measure.

Strongest bull case. Helios is a genuinely REIT-quality, long-dated, contracted asset that the market is under-crediting because it is wrapped in a crypto ticker; as Phase I/II rent flows and (critically) a second, non-CoreWeave hyperscale tenant signs, the data-center leg re-rates toward a diversified-landlord multiple, and the crypto franchise + treasury provide free optionality on the next crypto up-cycle.

Strongest bear case. GLXY is two correlated risks — CoreWeave’s balance sheet and the crypto cycle — with no durable moat under either, a controller who monetized $108M at the top, ~$1.2B/yr of external funding still required, and a valuation that already capitalizes the entire good asset. A CoreWeave wobble or a Bitcoin winter re-rates it violently.

The 3–5 assumptions that matter most: (1) CoreWeave remains solvent and pays for 15 years; (2) Helios Phases II–III are built on time and on budget with continued financing access; (3) a second, credit-worthy tenant eventually diversifies the single-tenant risk; (4) the crypto franchise’s normalized earnings are worth ~$3B; (5) Bitcoin does not enter a deep winter while the build is capital-hungry. Falsification: the bull breaks if the next tenant never signs and/or CoreWeave’s credit deteriorates; the bear breaks if a diversified IG tenant signs at Helios and Digital Assets sustains ~$200M+ operating pretax through a crypto flat patch.

12. Fact vs. Interpretation

13. Open Questions

1. CoreWeave credit. What are CoreWeave’s standalone liquidity, leverage, and contract-coverage metrics, and what happens to Helios cash flows (and the $1.4B facility) in a CoreWeave restructuring?
2. Helios economics precision. What is the actual stabilized cash gross margin and per-MW rent, and how are escalators/opex pass-throughs structured across Phases I–III?
3. Second tenant. Is there a credible pipeline for non-CoreWeave hyperscale tenants on the uncontracted ~800 MW+ of ERCOT-approved capacity, and on what timeline?
4. Net treasury exposure. What is the precise net directional crypto position and the ventures/PE book’s marks vs. realizable value?
5. Funding path. How much additional equity/debt is required to complete Phases II–III, and what dilution does that imply at current prices?
6. Normalized franchise earnings. Can Digital Assets sustain ~$150–200M+ of operating pretax through a crypto flat patch, or does it fade with volumes?
7. TRA drag. What is the expected cash cost of the Tax Receivable Agreement to public Class A holders over time?

14. What Must Be True

Bull case — what must be true:

• CoreWeave remains a paying, solvent tenant for the bulk of the 15-year term, and Helios Phases II–III energize on schedule and on budget with continued financing access.
• A second, credit-worthy hyperscale tenant signs on the uncontracted megawatts, converting single-tenant optionality into a diversified-landlord multiple.
• The Digital-Assets franchise sustains and grows normalized operating pretax (~$200M+), and the crypto cycle provides an up-leg that lifts the treasury.
• Falsification test: if, by end-2027, no non-CoreWeave tenant has signed and CoreWeave shows credit stress and Digital Assets pretax has faded back toward breakeven, the bull thesis is broken.

Bear case — what must be true:

• CoreWeave’s credit and/or AI-capex demand deteriorate, impairing the single-tenant Helios stream; or Bitcoin enters a deep winter while the ~$1.2B/yr build still needs external funding, forcing dilutive capital raises.
• The crypto franchise’s spreads compress as TradFi enters, and the treasury marks down.
• Falsification test: if Helios signs a diversified IG tenant, Phases II–III complete self-fundingly from lease cash flows, and Digital Assets holds ~$200M+ operating pretax through a flat crypto tape, the bear thesis is broken.

15. Source Appendix

See the separate Source Appendix (Appendix B in the combined report) for the full, dated, primarily-primary source list underpinning every claim above. Principal sources: Galaxy Digital FY2025 Form 10-K (filed 2026-02-26, period 2025-12-31) and Q1-2026 Form 10-Q (filed 2026-05-08); Form 4 insider filings (Aug-2025–Jun-2026); DEF 14A (2026); 8-K material-event filings (2025–2026); company press releases on Helios/CoreWeave (2025); aggregated fundamental data (cross-checked to filings); valuation-percentile and factor-model analysis; and public disclosures of peers Coinbase (COIN) and Core Scientific (CORZ).

This article expresses no investment recommendation and sets no price target, save for the clearly-labeled Claude's Take block, which is the author’s own opinion. It is general information, not investment advice.

APPENDIX A — Standard Diligence Questionnaire

Galaxy Digital Inc. (NASDAQ: GLXY) — as of 2026-07-11

Supplemental to the research memo. Labels: F=Fact, I=Interpretation, A=Assumption.

General

What thoughtful questions have other investors asked about this company? The debate centers on three axes: (1) Is GLXY an AI-data-center play or a crypto-beta stock? — the SOTP-vs-tape disconnect (a ~$19.4B sell-side SOTP vs. a ~$9.8B cap, while the factor model still clusters GLXY with COIN/MARA/BLOK). (2) How safe is the CoreWeave cash flow? — a 15-year lease is only as good as a sub-investment-grade, unprofitable single tenant (I). (3) How much dilution is left? — the company concedes it needs more capital to finish Helios, on top of a controller who sold $108M at $36 (F).

Cyclicality & Earnings Nature

• Cyclical high or low? Reported earnings are not on a stable cycle position — they are the crypto cycle. FY2024 (+$347M) was a crypto up-year; FY2025 (−$241M) a down-swing on Treasury marks; Q1-2026 a −$216M loss on a ~20% crypto drop (F). The operating franchise (Digital Assets pretax +$194M FY25) is at a cyclical high for volumes/spreads (I).
• External environment or internal actions? Overwhelmingly external (Bitcoin/crypto prices, trading volumes) for reported results; the one large internal value-creation lever is Helios (F/I).
• Revenue stability? Low. Transactional trading + mark-to-market dominate; the only annuity lines (asset-mgmt fees, staking/GK8, prospective Helios rent) are a thin recurring core (F).
• Market size / growth / geography? Global institutional crypto + AI-data-center; large and growing markets, but competitive and capital-attracting (I). US-centric assets (Helios in ERCOT), global counterparties.

Business Quality & Competitive Moat

• Industry getting more/less competitive? More. Crypto dealing/asset-mgmt is institutionalizing (spread compression, TradFi entry); AI data centers are a capital-cycle magnet (I).
• How profitable (ROIC/ROE)? ROE optically −19%/+24%/+50% (FY23–25) is leveraged crypto beta, not a quality signal; normalized operating returns are thin (F/I). ROIC ≈ meaningless for a mark-driven balance-sheet trader.
• Industry profitability / barriers? Crypto dealing: low barriers, many competitors (Coinbase, Binance, Kraken, Cumberland, B2C2, Wintermute, FalconX). Data centers: high barriers to power but many well-capitalized bidders for tenants (I).
• Easily understood? No — a three-segment, Up-C, mark-driven, crypto-plus-data-center hybrid whose GAAP is close to uninterpretable (I).
• Undermined by low-cost foreign labor? Not the relevant risk; the risks are crypto prices, tenant credit, and cost of capital (I).
• Do brands matter / switching costs? Novogratz’s brand carries deal flow (mild intangible); trade-level switching costs are ~zero for counterparties; real captivity exists only inside the signed CoreWeave lease (I).

Financial Condition & Balance Sheet

• Assets not fully on the balance sheet? The contracted >$15B Helios lease value is not on the balance sheet (upside); conversely, the TRA obligation and future funding needs are real drags (I).
• Off-balance-sheet liabilities? The Tax Receivable Agreement (85% of tax savings to insiders); ~$530M near-term data-center commitments; a ~$160M legal-settlement tail (F).
• How conservative is the accounting? Mixed — the lower-of-cost crypto model produces large two-way impairment/gain gross-ups (~$754M/$778M FY25) that obscure economics; segment disclosure is the honest lens (F/I).
• CapEx-hungry? Yes, acutely, right now — Helios drove $1.19B capex in FY25 and ~$1.2B/yr ongoing; the firm does not self-fund the build (F).

Capital Allocation & Management

• FCF generation / use / philosophy? FCF was −$1.5B in FY25 (Helios capex), covered by +$2.4B financing (F). Philosophy: reinvest aggressively across five arenas, fund with issuance; not a capital-returner (I).
• Significant acquisitions? GK8 (~$44M, 2023, cheap distressed custody buy) — good (F). The Helios site ($65M, 2022) — excellent (F).
• Buying back shares? A $200M authorization exists (Feb-2026) but is immaterial and pro-cyclical; GLXY is a massive net issuer (F).
• Issuing shares to insiders? Heavy SBC and RSU grants; more importantly, insiders (Novogratz $108M, Ferraro $27M) sold into the Oct-2025 raise (F).
• Compensation policy? Novogratz 2025 total ~$7.65M; bonus is discretionary with no formulaic ROIC/EPS grid; alignment rests on his ~$4.8B stake, undercut by the $108M sale (F/I).
• Motivations of management? Founder-controller (~49% vote) building a diversified digital-asset franchise; the $108M top-tick sale is the salient signal (F).

Valuation & Market Data

• ADR / MLP / K-1? Not an ADR. It is an Up-C structure (Class A common + Class B voting/LP units); public Class A holders own <50% of the operating partnership, with a TRA. No K-1 for Class A common holders (F).
• Dividend policy? No common dividend (F).
• How profitable? GAAP unprofitable TTM; normalized operating earning power ~$0.30–0.40/fully-diluted share (F/I).
• Net income vs. cash from operations diverging? Yes — positive GAAP NI in FY23/FY24 alongside negative operating cash flow every year; net income is dominated by non-cash marks (F).

Risks & Downside

• What causes the stock to decline? A CoreWeave credit/AI-capex wobble; a Bitcoin winter (impairs Treasury + forces dilutive raises); spread compression; failed Helios execution; further insider selling (I).
• Catastrophic-loss risk? A total loss is unlikely (ring-fenced Helios debt, matched trading book, modest net crypto), but a severe −50%+ drawdown is a live, correlated scenario (I).
• Total-loss chance? Low, but tail risks are fatter than a typical financial due to leverage + single-tenant + crypto beta (I).

Recent News & Events

• Environment changed recently? Yes — the AI-data-center re-rate (Helios/CoreWeave), the May-2025 NASDAQ listing, and the 2024–25 crypto-regulatory thaw all materially changed the story (F).
• Significant acquisitions / new markets / facilities? Helios build-out (ERCOT +830 MW Jan-2026); GalaxyOne retail launch (Oct-2025); K Wave Media advisory mandate (Jul-2025) (F).
• Accounting-policy changes? Ongoing evolution of crypto fair-value treatment; watch adoption of fair-value crypto accounting vs. the current impairment model (F/I).
• Leadership changes? Board re-elected May-2026; Novogratz remains controller/CEO (F).

APPENDIX B — Source Appendix

Galaxy Digital Inc. (NASDAQ: GLXY) — as of 2026-07-11

Primary sources first. Facts in the memo trace to these. Accessed July 2026 unless noted.

Primary — SEC filings (EDGAR, CIK 0001859392)

1. Form 10-K, FY2025 (filed 2026-02-26, period 2025-12-31) — glxy-20251231.htm. Segment note (Digital Assets / Data Centers / Treasury & Corporate pretax income), revenue/COGS gross-up, digital-asset impairment ($753.7M) & gain-on-derecognition ($778.2M), balance sheet, debt schedule (exchangeable notes, Deutsche Bank Helios term loan), Up-C/TRA disclosure, concentration disclosures, capex. https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001859392&type=10-K
2. Form 10-Q, Q1-2026 (filed 2026-05-08, period 2026-03-31) — glxy-20260331.htm. Q1-2026 net loss (−$216.3M), segment adjusted gross profit, Helios rent-commencement disclosure. https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001859392&type=10-Q
3. Form 4 insider filings (Aug-2025 – Jun-2026): Novogratz sale of 3.0M @ $36 (Oct-2025); Ferraro 750k @ $36; Deason open-market purchases (~$1.55M); director/officer routine dispositions. https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001859392&type=4
4. DEF 14A proxy (2026) — executive compensation (Novogratz ~$7.65M 2025 total; discretionary bonus), board nominees, ownership-guideline and TRA detail. https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001859392&type=DEF+14A
5. 8-K material events (2025–2026) — redomicile/NASDAQ listing (May-2025); June-2025 offering; $1.4B Helios term loan (Aug-2025); $460M placement + CoreWeave 526 MW (Oct-2025); $1.3B convertible (Oct-2025); ERCOT +830 MW (Jan-2026); Q4/FY25 & Q1-26 results; S-3ASR shelf (May-2026). https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001859392&type=8-K
6. S-1 / S-4 (2024–2025) — reorganization and US-listing registration; share structure and Up-C mechanics.

Primary — company disclosures

7. Galaxy press release — “Galaxy Completes Phase I of Its Helios Data Center Campus, Delivering 133 Megawatts of Critical IT Load to CoreWeave” (2025). Phase I 133 MW critical IT / ~200 MW gross; rent commencement Q2-2026. https://www.prnewswire.com/news-releases/galaxy-completes-phase-i-of-its-helios-data-center-campus-delivering-133-megawatts-of-critical-it-load-to-coreweave-302818664.html
8. Galaxy IR / investor materials — segment adjusted gross profit, AUM, counterparty count, assets-on-platform. https://www.galaxy.com

Secondary — trade press (Helios/CoreWeave)

9. Data Center Dynamics — “CoreWeave leases another 260MW capacity from Galaxy in Texas” (2025). https://www.datacenterdynamics.com/en/news/coreweave-leases-another-260mw-capacity-from-galaxy-in-texas/
10. The Block — “Galaxy delivers 133 MW of critical IT load to CoreWeave as Helios bitcoin mine turns AI hub” (2025). https://www.theblock.co/post/407396/galaxy-delivers-133-mw-of-critical-it-load-to-coreweave-as-helios-bitcoin-mine-turns-ai-hub
11. Crypto Briefing — “Galaxy Digital delivers first 200MW phase to CoreWeave under 15-year lease” (2025). 15-year term + two 5-year options; >$1B avg annual revenue. https://cryptobriefing.com/galaxy-digital-coreweave-200mw-helios-lease/

Quantitative data sources (cross-checked to filings)

12. Aggregated fundamental data (ROIC.ai) — income statement, balance sheet, enterprise value, profitability ratios, company profile (GLXY). Third-party aggregated; reconciled to the 10-K (filing authoritative).
13. Valuation-percentile analysis — GLXY’s own-history percentiles (P/B 77th, P/S 85th, composite 81st, 2026-07-10) derived from multi-year valuation data.
14. FactorsToday factor model — stock-loadings, stock-info (beta ~2.83, relative strength), related-stocks (nearest-neighbors), for GLXY. Third-party statistical estimates.
15. Public price history — 5-year OHLCV with moving averages, for the Five-Year Event Map (ATH $42.86 2025-10-21; low $16.84 2026-02-05).

Peer public disclosures (cross-read)

16. Core Scientific (NASDAQ: CORZ) public filings — the identical single-tenant CoreWeave data-center-landlord model; multiple discipline, single-tenant discount, capital-cycle framing.
17. Coinbase (NASDAQ: COIN) public filings — crypto-dealer/exchange quality-of-earnings and moat treatment; the “COIN-lite” comparison.

Aggregated data providers and factor models are third-party and are not primary; where they disagree with a filing, the filing governs.