Crypto & Digital Assets
PrivateA high-fee crypto-ETF toll-booth in structural runoff dressed up as a growth IPO — the cash cows (GBTC/ETHE) bleed faster than the long tail of cheap new products can replace the lost basis points, and DCG's no-economics control + Silbert litigation overhang make GRAY a melting-fee asset, not a crypto-beta proxy. WATCHING into the IPO; structurally BEARISH on the fee base.
Research
The verdict
A high-fee crypto-ETF toll-booth in structural runoff dressed up as a growth IPO — the cash cows (GBTC/ETHE) bleed faster than the long tail of cheap new products can replace the lost basis points, and DCG's no-economics control + Silbert litigation overhang make GRAY a melting-fee asset, not a crypto-beta proxy. WATCHING into the IPO; structurally BEARISH on the fee base.
Business model in plain terms. Grayscale packages crypto assets into regulated, brokerage-accessible fund wrappers and charges an annual management fee (accrued daily, paid in-kind out of the fund's crypto) on the assets it holds. It does not take principal risk on crypto; it is a fee-on-AUM toll-booth. Founded 2013, it is the oldest dedicated crypto asset manager and built its franchise on being the only compliant way for US brokerage and retirement money to hold bitcoin (2017–2023) before spot ETFs existed.
Key products. ~40+ products spanning ~45 tokens. The franchise splits into three tiers:
Main customers. Retail and institutional brokerage/RIA money accessing crypto through ordinary securities accounts (IRAs, 401ks, advisory platforms). No customer concentration in the B2B sense — the "customer" is the marginal ETF buyer, and that buyer is fee-sensitive and rotating to cheaper rivals.
Main suppliers. Crypto custodians (historically Coinbase Custody as the primary custodian for the trusts), index providers (CoinDesk Indices — note: also a DCG sibling), market makers / authorized participants, and the exchanges (NYSE Arca). Custody is effectively single-source-ish (Coinbase), a chokepoint (Lens 2).
Main competitors. BlackRock (IBIT), Fidelity (FBTC), Bitwise (BITB / BSOL), ARK 21Shares (ARKB), VanEck/Franklin/etc. in spot crypto ETFs; Galaxy Digital and Coinbase Asset Management in the broader digital-asset-manager category.
Contract / fee structure. Revenue is recurring, AUM-linked, daily-accruing — high quality if AUM is stable, but Grayscale's is not stable: it is (a) levered to crypto price (AUM rises/falls with BTC/ETH) and (b) levered to fee compression and redemptions. No take-or-pay, no lock-ups — ETF money can leave on any trading day, and it is leaving (Lens 4/5).
Map (each named stakeholder along the chain):
Upstream inputs →
→ Grayscale (the manager) — wraps inventory in a fund, accrues the fee, handles creation/redemption with APs.
→ Distribution / market plumbing:
→ End customer: retail + institutional brokerage holders.
Chokepoints & single-source dependencies:
The honest answer: the historical moat is gone, and the residual moats are thin.
Bargaining power. Over customers: weak and weakening (they have a cheaper substitute one click away). Over suppliers: weak (Coinbase custody, CoinDesk index — and the index is a related party, so "bargaining" is internal politics). Net: Grayscale needs its customers far more than they need Grayscale. That is the inverse of a moat.
No segments.csv (empty — private). Segmentation is by product and by asset, from the S-1 disclosures via outlets:
| Product | Asset | ~AUM | Fee | ~Annual revenue | Notes | Source |
|---|---|---|---|---|---|---|
| GBTC | Bitcoin | ~$15–17.3B | 1.5% | ~$260M | The cash cow; ~$25B of outflows since Jan-24 conversion | |
| ETHE | Ethereum | ~$3.4B | 2.5% | ~$85M | Second cash cow; ~$4.8B outflows since Jul-24 conversion | |
| Mini trusts (BTC/ETH) | BTC/ETH | small | 0.15% | de minimis | Defensive clones to retain rotating assets | |
| GDLC / GSOL / single-token / staking / covered-call | multi/SOL/etc. | the rest (~$12–14B of ~$35B) | mixed, low | the rest (~$80M) | The "growth" tail |
The single most important number in the whole dossier (concentration):
GBTC + ETHE = ~70% of AUM but ~88% of 9M-2025 revenue (and ~93% of FY2024 revenue).
i.e. the ~$345M thrown off by the two legacy trusts vs ~$425M total run-rate. The "diversified 45-token platform" is, on a revenue basis, two declining high-fee products with a rounding-error tail.
Geography: primarily US (US-listed ETFs, US brokerage distribution). Not broken out; treat as ~US. n/a — geographic split not disclosed.
Trend (decelerating, and why): average AUM slipped to $30.6B (9M-2025) from $31.8B (9M-2024) despite a higher crypto price tape over much of the period — meaning share loss, not market shrinkage, is doing the damage; and the weighted-average fee fell to 1.39% from 1.67%. Both the price (fee) and the volume (AUM) are falling at once — the textbook definition of a business in structural decline.
No financials.csv rows (empty — private). Hard P&L from the S-1 (via outlets — all ``):
| Metric | FY2023 | FY2024 | 9M-2024 | 9M-2025 |
|---|---|---|---|---|
| Revenue | $512.7M | $506.2M | $397.9M | $318.7M |
| Net income | $325M | $282.1M | $223.7M | $203.3M |
| Weighted-avg mgmt fee | n/a | ~1.67% (9M-24) | 1.67% | 1.39% |
| Average AUM | n/a | n/a | $31.8B | $30.6B |
Read of the print:
No transcripts/ (empty — private, no earnings calls yet). +private overlay: substitute founder/CEO interviews & public positioning.
| Company | Ticker | Mkt cap (USD) | Business | P/E | Notes | Source |
|---|---|---|---|---|---|---|
| Grayscale | GRAY (pending) | see valuation conflict below | Crypto fund mgr (fee-on-AUM) | n/a (private; ~64% net margin, ~$270M NI run-rate) | The subject | |
| Coinbase | COIN | ~$38.8B (2026-06) | Exchange + custody + asset mgmt | positive (profitable) — exact P/E n/a | Diversified, not a pure comp | |
| Galaxy Digital | GLXY | ~$8.4–13.3B (2026-06) | Trading + asset mgmt + data-center pivot | negative (EPS TTM −1.73; Q1-26 net loss $216M) | Unprofitable; AUM+stake ~$8B | |
| Bitwise | private | n/a | Crypto ETF sponsor | n/a | ~67% of SOL-ETF AUM | |
| BlackRock (IBIT only) | BLK | n/a (IBIT is a product) | The fee-war aggressor | n/a | ~49–60% of spot-BTC ETF AUM at 0.12–0.25% |
The valuation conflict — surfaced, not resolved (per provenance rules):
Two incompatible numbers circulate. (A) Analyst chatter pegs a "$30–33B valuation" — but this appears to conflate AUM (~$35B) with enterprise value, which is a category error for a fee manager (you don't capitalize the clients' assets, you capitalize the fee stream). (B) A P/S-based estimate puts market cap at ~$1.5B (at 3× sales) to ~$3.4B (at 7× sales) on ~$487M revenue. These differ by ~10×. The (B) framing is the analytically correct order of magnitude for a declining ~$270M-NI manager; (A) is an AUM headline masquerading as a valuation. If GRAY prices anywhere near the (A) "$30B" framing, that is the short. I deliberately do not average these.
Anchoring sanity check ``: A growing high-quality asset manager trades ~15–25× earnings; a declining one trades ~8–12×. At ~$270M run-rate NI × 10× = ~$2.7B; × 15× (if you believe the diversification story arrests the decline) = ~$4B. That triangulates to the (B) range, far below any "$30B." `]
Grayscale-the-equity isn't trading yet, so the catalyst history is read off (i) the GBTC discount/flow tape and (ii) crypto-beta:
What the pattern reveals: the market reacts to (1) the fee gap / flows, (2) crypto price (AUM beta), (3) DCG/Silbert legal headlines. It does not reward product-count press releases. The stock (once listed) will be a levered, fee-decaying play on the crypto tape — more downside beta than upside, because rising BTC lifts AUM but doesn't stop the rotation to cheaper rivals, while falling BTC hits AUM and accelerates fee-sensitive redemptions.
No financials.csv/filings on disk; the S-1 supplies the disclosures (via outlets). Forensic flags:
Regulatory findings (required sub-section):
regulatory/regulatory-findings.md (generated 2026-06-30) reports 0 SEC findings — Grayscale has no CIK and is not an SEC registrant for enforcement-search purposes, so no EDGAR LR/AAER hits are possible."Grayscale Investments" (FTC OR DOJ OR FDA OR CFPB OR consent decree OR settlement OR fine OR penalty) enforcement): No material enforcement action against Grayscale Investments itself surfaced. The material legal matters attach to parent DCG and Barry Silbert (NY AG civil fraud — see flag #4 above), not to Grayscale. Grayscale's own most consequential legal event was as a plaintiff (Grayscale v. SEC, won Aug-2023).n/a — Grayscale is private and files no 10-K. No 10-K exists on disk or at EDGAR for the operating company; none is invented. (The S-1 will carry a "Legal Proceedings" section; not separately fetched here — flagged as an open item.)+private → fee-on-AUM model + path-to-tradeable)No guidance.csv (empty). Build it bottom-up from S-1 actuals; output ``, arithmetic shown. The driver is AUM × blended fee × ~64% net margin. Modeling NI (no public share count yet — EPS deferred to IPO terms).
Anchor (actuals): 9M-2025 revenue $318.7M → ~$425M run-rate; NI ~$203M 9M → ~$270M run-rate; blended fee 1.39% and falling; avg AUM ~$30.6B.
| Scenario | Avg AUM path | Blended fee path | Implied revenue | Implied NI (~64% margin) | Logic |
|---|---|---|---|---|---|
| Bear | $30.6B → ~$22B (−10–15%/yr: continued GBTC/ETHE runoff + soft crypto tape) | 1.39% → ~1.05% (mix-shift accelerates) | ~$145M | The structural-decline base rate continues; crypto chops/falls. | |
| Base | $30.6B roughly flat → ~$30B (new products + crypto price offset legacy runoff) | 1.39% → ~1.15% (mix-shift continues, just slower) | ~$220M | Diversification slows but doesn't stop the bleed; revenue keeps drifting down. | |
| Bull | $30.6B → ~$40B (crypto bull-run + staking/income products win net new) | 1.15% → ~1.20% (yield products carry higher fees) | ~$305M | Crypto rips and the breadth/yield strategy compounds faster than GBTC bleeds. Best case ≈ flat to 2024. |
The tell: even the bull case only gets revenue back to roughly its 2024 level — this is structurally a no-growth-to-declining earnings stream. You are underwriting how fast it shrinks, not how fast it grows. A fee manager whose price and volume both fall is worth a low single-digit P/AUM and a single-digit-to-low-teens P/E — which is why the (B) ~$1.5–4B valuation framing (Lens 7) is the right order of magnitude and the "$30B" headline is a trap.
Path-to-tradeable (the +private payoff lens): Grayscale is further along than any other private on the crypto frontier — S-1 publicly on file since 2025-11-13 (after a July-2025 confidential DRS), underwriters lined up (Morgan Stanley, BofA, Jefferies, Cantor), ticker GRAY reserved, Up-C structure set, $250M pre-IPO preferred closed, Silbert board cleanup done. The only gates left are SEC effectiveness + pricing + market-window — it can list on any clear window in 2026. This is the single most important fact for actioning the name: it is about to become directly tradeable, so the analysis converts straight into a long/short/avoid view on GRAY at its IPO price.
No
private-watch.jsonentry exists for grayscale, and wave boundaries forbid editing it. Flagged for Connor: add grayscale toprivate-watch.jsonwithstage: S-1-filed,ipo_readiness: imminent,catalyst: GRAY NYSE pricing (2026), anddossier: companies/grayscale/deep-dive-2026-06-30.md— this is exactly the dossier-warm IPO-proximity name that ledger is built to surface.
Forecast log: Skipped per --watchlist rule (no forecast.ts create in the breadth loop). If actioned, the scoreable forecast would be on the IPO outcome, e.g. "GRAY prices at an implied equity value ≤ $6B" (p≈0.65 ) — far below the "$30B" AUM-conflated headline — to calibrate the "melting-fee, not crypto-beta" thesis.
Bull case (narrative). Grayscale is the incumbent brand of crypto asset management at the exact moment the asset class goes fully mainstream and regulated. The fee war on plain-vanilla BTC is lost, yes — but the next leg of crypto AUM is staking yield, income (covered calls), and multi-asset baskets, and Grayscale is first to market in nearly all of them (first US multi-asset ETP, first ETH staking-reward distribution, a covered-call income suite). Those products carry higher fees and stickier holders than commodity BTC exposure. Layer in a rising crypto tape lifting AUM, a 64% net margin that converts every retained dollar to profit, and a clean post-IPO currency to acquire/consolidate smaller sponsors — and a sub-$4B IPO of a ~$270M-NI franchise could re-rate as the decline arrests. The contrarian bull: the market is pricing GBTC's funeral and ignoring that Grayscale is quietly winning the long-tail product race.
Bear case (the base rate — and the stronger case). Three things can permanently impair the franchise: (1) the cash cows never stop bleeding — GBTC's 1.5% fee is indefensible vs IBIT's 0.12–0.25%, and tax lock-in only delays the rotation while no new money ever enters at 1.5%; (2) the long tail can't carry the load — even the bull case only claws revenue back to 2024 levels, because 0.15% mini-trusts and small single-token funds need enormous AUM to replace a basis-point of GBTC; (3) the controller — DCG/Silbert's super-voting control + live NY AG fraud litigation is a permanent governance discount, and the Up-C secondary tells you insiders are cashing out at the top, not investing in the turnaround. Expectations baked into any "$30B" framing are detached from a declining-earnings reality. Multiples on a shrinking 64%-margin manager belong in the high-single-digits — anything more is the IPO underwriters selling the AUM headline.
Pre-mortem (18 months out, thesis broke). It's late 2027. GRAY priced into a crypto bounce at a flattering multiple; then BTC rolled over, AUM fell 25%, GBTC outflows re-accelerated (the fee-sensitive marginal holder always leaves on weakness), the blended fee broke below 1.1%, NI fell toward $150M, and the stock is down 50%+ from IPO. The staking/income products grew but off a tiny base and at fees too low to matter. A DCG-litigation headline periodically caps any rally. The post-mortem line: we bought the brand and the AUM headline and ignored that the unit economics (fee × retention) were structurally negative.
Are multiples too high? At the "$30B" AUM-anchored framing — emphatically yes. At the P/S-derived ~$1.5–4B — defensible to cheap, but only if you believe diversification arrests the decline. The entire investment question is the IPO price.
Contrarian view (what the market refuses to see). Both sides. Bulls refuse to see that a 64% margin on a melting fee base is a liability multiplier, not a quality signal — it means earnings fall as fast as revenue. Bears refuse to see that Grayscale's product-velocity edge in staking/income/multi-asset is genuinely real and under-modeled — it just isn't yet big enough to matter. The synthesis: GRAY is a value/melting-ice-cube stock, not a crypto-growth stock — and it will almost certainly be marketed as the latter. The edge is refusing the framing.
Dismantling the bull case:
A toll-road compounder mispriced as a disruption victim — 60%+ margins and 16% top-line growth are intact while the market discounts a debit antitrust loss and a stablecoin bypass that the numbers (cross-border +17%, $7B stablecoin run-rate is on-network, not against it) do not yet support; structurally BULLISH, but the DOJ debit case and the interchange settlement's final approval are real, dateable downside.
A toll road on global consumption priced like a bond proxy — the moat is intact and value-added services are compounding at 2x the network, but the stock has de-rated to ~24x forward because the market is (rightly) pricing two live structural threats — stablecoin disintermediation and large-issuer network defection (Capital One/Discover) — that bulls keep waving away. WATCHING, lean BULLISH on weakness.