Galaxy Digital GLXY

Reference table

1. What it is

Galaxy Digital Inc. (Nasdaq: GLXY) is a digital-asset and data-center-infrastructure company operating through three segments: Digital Assets (OTC spot and derivatives trading, lending, structured products, M&A advisory and capital-markets services, asset management, plus blockchain infrastructure — staking, tokenization, custodial technology), Data Centers (the Helios infrastructure assets), and Treasury and Corporate (a proprietary portfolio of digital assets, ventures, private-equity and fund investments, plus bitcoin mining operations). It also runs GalaxyOne, a retail fintech platform bridging traditional and digital markets. Founded 2018, headquartered at 300 Vesey Street, New York; CEO Michael Novogratz; ~700 employees. The Nasdaq listing is young — US IPO date 2025-05-16 (barely 14 months of US trading history). The balance sheet carries a $1.11B minority interest, reflecting the partnership/Up-C-style structure beneath the public entity.

Revenue mix — a data gap to flag honestly: the FMP feed supplies no product-segment or geographic revenue split for GLXY (both segment blocks are empty). What we can say from the statements: FY2025 "revenue" of $61.36B against cost of revenue of $60.21B (gross profit $1.15B, a ~1.9% gross margin) is the signature of gross trading-flow accounting — the top line is dominated by the notional value of digital-asset sales, not fee-like revenue. Judge this company on gross profit, operating income, and cash flow, never on the headline revenue or the (meaningless) 0.14× P/S.

The strategic story: a crypto merchant bank whose real equity kicker is the conversion of Helios from bitcoin mining to AI/HPC data-center capacity — moving revenue from pure crypto-cycle beta toward contracted infrastructure.

2. The expert thesis (traceable)

No expert-panel coverage — this note is fundamentals-driven. The Synthos KB contains zero claims on GLXY across zero voices (grep of the labeled claim store, 2026-07-06). That is the honest house standard for a screen-surfaced name: no borrowed conviction, no paraphrased "street chatter" dressed up as expert signal. Everything in this note reconciles to the FMP data file (statements, estimates, targets, technicals, insider filings) and to nothing else.

Two consequences, stated plainly:

• Conviction is capped at Low. Without a single high-skill voice underwriting the thesis, this cannot be a conviction-track name regardless of how interesting the fundamentals look. The verdict below is deliberately conservative.
• There is no independent bear thesis either. The §3 bear case is built from the fundamentals (cash burn, dilution, crypto-cycle exposure), not from a countervailing expert — treat its confidence accordingly.

If the panel ever picks up GLXY (Novogratz is a prominent public voice; coverage could appear), this note gets re-versioned with reconciled claim IDs.

3. Synthos scores & the Bull / Base / Bear cases

The one-glance judgment — three scores, 0–10, each anchored to real metrics (not probabilities we can't honestly calibrate):

The three cases (our own scenario model — assumptions shown; each target is a ~12–18-month fair value). We deliberately do not attach probabilities: the base case is the expected path; the cases bound the range; the scores summarize them.

Synthos fair value = the base case, ~$31 (+22%), full range $15–$46. Two honesty flags on the anchor: (1) with negative TTM earnings and negative FCF there is no clean cash-flow anchor — the base case leans on the Street's $33.67 consensus (only ~11 analysts, high $40 / low $24) plus a book-value cross-check, and we label that dependence rather than dress it up as a DCF; (2) the FMP forward-estimate block is internally inconsistent (2026E "EBITDA" of ~$28B against near-zero net income is an artifact of gross-revenue accounting) — we use only the EPS and revenue lines from it, and even those thin out to 1–5 analysts beyond 2026. The 3.1× bull/bear ratio is the honest signal: this is a high-variance, cycle-driven name. This is a tracked call — the Forecaster Scorecard grades it once it matures.

4. Exponential Potential

Synthos separates compounders (durable high returns on capital) from exponentials (accelerating, multi-baggers-from-here). GLXY is a conditional exponential — the optionality is real, the acceleration is not yet:

• Two live TAMs, one small cap. At $8.42B, GLXY holds an institutional digital-asset franchise (OTC trading, lending, staking, tokenization, custody tech, asset management) and the Helios data-center campus pivoting from bitcoin mining to AI/HPC. Either leg working at scale re-rates the whole company; both working is the multibagger path.
• The Helios pivot is the differentiated leg. PP&E jumped from $237M (FY24) to $1.42B (FY25) and FY25 capex was $1.19B — the buildout is visibly happening on the balance sheet, not just in the deck. Contracted data-center revenue would be the first non-crypto-cycle cash flow in the company's history. (The data file provides no lease/tenant detail — we can verify the spend, not the contracts.)
• The acceleration test currently fails. Quarterly gross revenue ran $12.98B → $9.06B → $29.22B → $10.22B → $10.21B over the last five quarters — that is volatility, not acceleration — and consensus has 2026 gross revenue down (~$46.6B vs $63.1B 2025E) with EPS still negative (−$0.32) before a thin-coverage swing to +$0.26 in 2027.
• Reinvestment is externally funded. Unlike a true exponential compounding its own cash, GLXY's runway is financed by debt ($1.66B issued FY25) and equity (+59% share count in five quarters). The option is being paid for with your dilution.

Exponential Potential: 7/10 — High, but conditional. The right way to own this (when the setup improves) is as a small option on crypto-cycle beta plus Helios execution — not as a compounder position.

5. Financials (real numbers — FMP annual/quarterly)

• Revenue (gross trading flow — read with care): FY25 $61.36B, +40.2% (FY24 $43.76B). Cost of revenue $60.21B → gross profit $1.15B, a 1.9% gross margin. FY23 and earlier show $0 revenue in this feed (pre-restructuring/Up-C presentation) — the multi-year revenue history is not comparable and we don't pretend otherwise.
• Quarterly trajectory (the volatility is the story): Q1'25 $12.98B (EPS −$2.31) → Q2'25 $9.06B (+$0.10) → Q3'25 $29.22B (+$1.19, operating income $756M) → Q4'25 $10.22B (−$1.03, operating income −$224M) → Q1'26 $10.21B (−$0.49, operating income +$67M; missed the −$0.44 estimate). One monster crypto quarter, then two losses.
• FY25 P&L: operating income $575M, EBITDA $609M — but −$846M of other income/expenses below the line drove pre-tax to −$271M and net income to −$84.9M (EPS −$0.53). TTM: net margin −0.11%, EPS −$0.35.
• Cash flow — the crux: FY25 operating cash flow −$256M, capex −$1.19B (Helios buildout), FCF −$1.45B. Funded by $1.66B net debt issuance + $851M equity issuance; $49M of distributions also paid. TTM FCF yield −23.4%.
• Balance sheet: cash $1.45B + short-term investments $5.22B = $6.67B liquid; total debt $5.33B (of which $2.84B short-term), FMP net debt $3.88B, net-debt/EBITDA 2.30×. Total assets $11.35B include $3.59B goodwill + intangibles (under crypto-industry accounting this bucket typically includes digital-asset holdings — the feed doesn't break it out). Stockholders' equity $1.92B plus $1.11B minority interest; tangible book per share −$0.14; book value per share $14.47; current ratio 1.70×.
• Dilution: weighted shares 120.8M (FY24) → 159.2M (FY25) → 192.1M (Q1'26) — +59% in five quarters. Every forward per-share number must be read against a still-rising denominator.

6. Valuation — priced in or room?

There is no honest earnings multiple here: TTM P/E is negative (−110× on a −$0.35 EPS), and the headline P/S of 0.14× is meaningless on gross trading revenue. What the data supports: P/B 2.69× on $14.47 book (and infinite on negative tangible book), EV/EBITDA 11.3× TTM on volatile EBITDA, FCF yield −23%. FMP's letter rating is C+ (overall 2/5; ROE and P/E score 1/5 — the DCF sub-score of 5/5 says less about value than about how badly mechanical DCFs handle gross-revenue financials). Forward: consensus 2026 EPS is −$0.32 (9 analysts), swinging to +$0.26 in 2027 (5 analysts, range −$0.21 to +$0.57) and ~$0.19 in 2028–29 (one analyst — effectively no coverage). Even taking 2027's +$0.26 at face value, $25.40 is ~98× a number nobody has confidence in — the market is not paying for earnings, it is paying ~2.7× book for the franchise plus the Helios option. Street targets (context): consensus $33.67, median $35.50, high $40 / low $24; 9 Buy · 2 Hold. Our base case ($31) sits just below consensus, and we flag openly that with no cash-flow anchor the consensus itself is doing real work in that number. Not a value buy; an optionality-at-a-fair-ish-price situation where the entry setup matters more than the multiple.

7. Technicals (from the tech block)

• Trend: broken. $25.40 sits below both the 50-DMA ($29.39) and the 200-DMA ($27.83), and the 50 has crossed under the 200 (death-cross posture). MACD −0.95 (negative).
• Location: −41% off the 52-week high ($42.86) — also the max drawdown from peak — and +51% off the 52-week low ($16.84). (The quote block shows a slightly wider $16.43–$45.92 range; we use the tech block consistently and flag the feed mismatch.)
• Momentum: RSI(14) 22.4 — deeply oversold. This is capitulation territory, which is exactly where Watch names earn their trigger reviews.
• Relative strength: +44.0% 3-mo vs SPY +14.6% / QQQ +23.6% — but only +14.3% 12-mo vs SPY +21.1% and QQQ +31.2%, and +13.6% 6-mo. A violent 3-month rally that has already given much of itself back; 12-month relative strength is negative vs both benchmarks.
• Read: technicals are weak — downtrend below both moving averages with an oversold flush in progress. For a 3.61-beta name, catching this knife is a trade, not an investment; the disciplined play is to let the base form (reclaim of the 200-DMA ~$27.8) or let the flush complete toward ~$20 before the Watch upgrade discussion.

8. Moat & competitive position

Galaxy's edge is regulated-adjacent institutional plumbing in a trust-scarce industry: an OTC desk, lending book, staking/custody technology and asset-management arm that institutions can face without touching an offshore exchange, fronted by one of crypto's most recognizable principals. That franchise is real but narrow and cyclical — it competes with Coinbase's institutional arm, traditional prime brokers edging in, and pure-play market makers, and its revenue rises and falls with crypto volumes it does not control. The Helios data-center leg competes on powered land and speed-to-energization, a genuinely scarce asset in the AI buildout — but there GLXY is a small new entrant against hyperscale-scale developers, and the feed gives us no contracted-tenant detail to underwrite. ROIC 11.4% TTM (FMP) against ROE −3.8% tells you the operating assets work while the total entity doesn't yet.

Peer set (FMP-supplied, market cap): a mixed and only half-relevant list — Franklin Resources $17.9B, Ally Financial $14.2B, Evercore $13.8B, Stifel $11.5B, Jefferies $11.2B, Houlihan Lokey $9.8B, Cipher Mining $8.9B, Riot Platforms $8.6B, XP $8.6B. The traditional brokers/advisors are poor comps; the honest cohort is the miners-pivoting-to-AI-datacenters (CIFR, RIOT — both within $0.5B of GLXY's cap) plus Coinbase/Circle-type crypto financials, which are not in the supplied set — a data caveat. Against the miner cohort, GLXY brings a real financial-services franchise the miners lack; against the financials, it brings the Helios asset they lack.

9. Management, capital allocation & guidance

• Capital allocation: all-in on the Helios buildout — FY25 capex $1.19B (vs ~$0 in FY24 as presented), funded by $1.66B net new debt and $851M of new equity, alongside $49M of distributions. No buyback, no common dividend (lastDividend 0). This is a deliberate bet-the-growth posture; the cost is 59% share-count growth in five quarters and a levered balance sheet before the asset produces.
• Insider activity (Form 4s in the file): the pattern is selling into strength in June 2026 — CAO Robert Rico exercised 24,000 options at $4.83 and sold 38,000 shares at $31.62–$34.22 across 2026-06-09/15 (classic exercise-and-sell, but a >6× spread taken off the table near the highs). CEO Michael Novogratz and President/CIO Christopher Ferraro show only F-InKind dispositions (4,683 and 3,891 shares at $29.58 on 2026-06-01 — routine tax-withholding on vesting, not discretionary sales). No open-market buys in the window. Mildly negative tell, not a red flag.
• Management guidance: not available in our data — the FMP feed carries no guidance block or transcript for GLXY, and with zero KB claims there is no ingested management commentary to half-weight. We flag the gap rather than paraphrase press coverage. What the calendar shows: Q1'26 missed (−$0.49 vs −$0.44 est) after Q4'25 beat (−$1.03 vs −$1.24 est) — estimate quality on this name is poor in both directions.

10. Catalysts & what to watch

• Next earnings: 2026-08-04 (Q2 2026; Street EPS −$0.05, revenue est ~$11.0B). The key lines: Data Centers segment revenue (is Helios generating yet?), the capex run-rate, and whether another equity raise is signaled.
• Helios lease/energization announcements: contracted AI/HPC capacity with credible counterparties is the single most valuable de-risking event — it converts crypto-cycle beta into infrastructure cash flow.
• Crypto cycle: trading, lending, staking and the treasury book are all long the same factor; a sustained digital-asset rally is the fastest path to the bull case (Q3'25 showed what one hot quarter does: +$1.19 EPS).
• Financing events: with FCF at −$1.45B and $2.84B of short-term debt on the book, any new issuance (equity or debt), refinancing terms, or asset sales are thesis-relevant.
• Technical repair: a reclaim of the 200-DMA (~$27.8) on volume, or a completed flush toward ~$20, are the two setups that would trigger a re-review of the Watch verdict.

Thesis tripwires (what would change the call): upgrade triggers — Helios revenue turning on with disclosed contracted tenants, two consecutive FCF-positive quarters, or a washout entry near ~$20 with crypto fundamentals intact. Downgrade-to-avoid triggers — another large dilutive raise at depressed prices, short-term debt refinanced on punitive terms, or a crypto drawdown forcing treasury-book liquidation.

11. Key risks

• Crypto-cycle concentration (the dominant risk): trading volumes, lending demand, staking yields and the treasury/venture book are all long the same underlying factor. A digital-asset bear market hits every segment at once — while the Helios bill is still being paid.
• Cash burn + financing dependence: −$256M operating cash flow and −$1.45B FCF in FY25, plugged with debt and equity. If capital markets close to crypto names, the buildout slows or the dilution gets worse.
• Dilution: +59% share count in five quarters, and the 2027+ per-share estimates rest on a denominator that is still growing.
• Leverage & liquidity structure: $5.33B total debt with $2.84B current, net-debt/EBITDA 2.3× on volatile EBITDA, financial leverage 5.5×, tangible book negative. The $6.67B liquid-asset pool is the counterweight — but much of it sits against the trading/treasury book, not free cash.
• Execution risk at Helios: a small new entrant building hyperscale-class infrastructure; construction, energization and tenant-credit risk, with no contracted-revenue disclosure in our data to verify progress.
• Volatility itself: beta 3.61 and a demonstrated −41% drawdown inside a year — position-sizing risk is a first-order risk here, not a footnote.
• Coverage thinness: ~11 analysts on ratings, 1–5 on out-year EPS; the consensus our base case leans on is itself fragile.
• Accounting opacity: gross-revenue presentation, a large undifferentiated intangibles bucket, minority interest, and empty segment data in the feed — the statements require more trust than a plain operating company's.

12. Verdict, position sizing & monitoring

Watch. Galaxy Digital is one of the more interesting optionality stacks in the momentum pool — a real institutional crypto franchise plus a verifiable (on the balance sheet, at least) AI-data-center buildout, at an $8.4B cap, with the Street 33% above the last price. But the disqualifiers for entry today are equally real: a TTM loss, −$1.45B of FY25 free cash flow funded by 59% dilution and $1.7B of new debt, a 3.61 beta, zero expert-panel coverage, and a broken tape sitting below both moving averages mid-flush. The house standard is conservative on screen-surfaced names, and this one earns it: the option is worth watching, not yet worth paying for.

• Sizing: none today. If an upgrade trigger fires (§10), initial sizing is tactical, ≤1% of the flagship — a 3.61-beta, negative-FCF name must be sized so a 50%+ drawdown is a non-event.
• Entry discipline: the two setups that reopen the discussion are a washout toward ~$20 (crypto fundamentals intact) or a 200-DMA reclaim (~$27.8) with Helios revenue disclosure — price alone isn't enough; the second condition is proof the non-crypto leg exists.
• Monitoring: re-underwrite on the §10 tripwires; formal re-score each earnings print (next 2026-08-04). This verdict is logged as a tracked Synthos call as of 2026-07-06 at $25.40.
• Single biggest risk: a crypto drawdown landing mid-buildout — cash burn, short-term debt, and dilution dependence with no internal cushion.

Provenance & disclosures

• Traceability: 0 KB claims, 0 expert voices — a grep of the labeled claim store on 2026-07-06 returned nothing for GLXY, and per house standard this note therefore cites no expert and carries a Low conviction cap. Fabricated conviction is structurally impossible (claim-ID reconciliation); kb_net_conviction and kb_claims date are null rather than invented.
• Data as-of: fundamentals 2026-03-31 (Q1 2026, filed 2026-05-08) · estimates, targets, prices, insider filings 2026-07-06 (FMP pull). Forward figures are analyst consensus, labeled as estimates.
• Revenue-accounting caveat: GLXY's headline revenue is gross digital-asset trading flow (~2% gross margin); P/S-style metrics are not meaningful and are not used in the valuation. Pre-2024 revenue lines in the feed show $0 and are not comparable.
• Estimate-quality caveat: the FMP forward EBITDA/SG&A lines are internally inconsistent (artifacts of gross-revenue aggregation) and were excluded; EPS/revenue estimates rest on 1–9 analysts depending on year.
• Segment caveat: product and geographic segment blocks are empty in the feed — Digital Assets vs Data Centers revenue split could not be verified from our data.
• Guidance caveat: no management guidance or transcript is available in our data; none is paraphrased.
• Peer caveat: the FMP-supplied peer list is dominated by traditional brokers/asset managers; the honest comparators (CIFR, RIOT, and crypto financials like Coinbase not in the list) should frame any relative-value work.
• Not investment advice. Independent research, educational and informational only, never personalized. Hypothetical/forward figures are labeled; the only performance numbers Synthos will headline are the live, real-money Flagship's.
• Version: 2026-07-06. Prior versions available via the deep-dive version dropdown ("based on the info at the time").