Crypto & Digital Assets
A sub-1%-share, structurally subscale crypto exchange whose core trading revenue is already flat — kept alive by recurring related-party rescue financing from its own founders; the ~85% post-IPO drawdown is the market correctly re-rating a $3.3B IPO to a ~$0.5B going concern, and nothing in the numbers says the floor is in.
Research
The verdict
A sub-1%-share, structurally subscale crypto exchange whose core trading revenue is already flat — kept alive by recurring related-party rescue financing from its own founders; the ~85% post-IPO drawdown is the market correctly re-rating a $3.3B IPO to a ~$0.5B going concern, and nothing in the numbers says the floor is in.
Gemini is a US-regulated, retail-and-institutional crypto exchange, custodian, and stablecoin issuer founded in 2014 by Tyler and Cameron Winklevoss. It went public on Nasdaq (GEMI) on 12 Sep 2025 at a $28.00 IPO price; the holding company (Gemini Space Station, Inc., Nevada, formed 4 Feb 2025) sits atop an Up-C structure controlling Gemini Space Station, LLC ``.
How it makes money — five engines, but one dominant:
Customers — 601K Monthly Transacting Users and ~10,000 institutions across 60+ countries as of 31 Dec 2025 . Heavily US-weighted, and getting more so: in Feb 2026 management announced an exit from the UK, EU, and Australia, leaving only the US + Singapore ``.
Suppliers / counterparties — Mastercard + WebBank (credit card), Ripple (co-brand + a credit facility funding card receivables), NYDIG and Galaxy (historical lending/collateral), and — critically — Winklevoss Capital Fund, LLC (WCF), the founders' own fund, as the lender of last resort (see Lens 9/10).
Contract structure — fee-based and non-recurring: transaction revenue is pay-per-trade with no take-or-pay floor, so revenue is a direct derivative of crypto trading volume and prices. The company says so explicitly: "A significant amount of the trading volume on our platform is derived from a relatively small number of institutional market makers" — about half of total volume ``. That is the single most important structural fact about this business: concentrated, volatile, and un-contracted.
Crypto exchanges don't have a physical supply chain; the analogue is the liquidity + settlement + fiat-rails + financing stack. Named stakeholders end-to-end:
Chokepoints / single-source dependencies: (1) the market-maker concentration (~50% of volume); (2) related-party financing from WCF — without it, the historical capital-reserve and collateral obligations could not have been met; (3) third-party network dependence (Mastercard/WebBank) for the fastest-growing services line. Names are present — this lens passes.
Gemini's self-described moat is "the most trusted brand in crypto": early NYDFS license, GUSD as one of the first regulated stablecoins, first regulated exchange to list ether (2016), SOC/ISO certifications, and a vertically integrated "superapp" ``. That is a real but shallow and non-exclusive moat:
Verdict on moat: narrow-to-none. The vertically integrated superapp story is a strategy, not a moat; at <1% share it has neither the cost advantage of scale nor a defensible switching cost.
No GAAP operating segments are reported (single reportable segment), but revenue disaggregation `` is the segment view. All figures FY ($000s):
| Revenue line | FY2025 | FY2024 | FY2023 | FY25 trend |
|---|---|---|---|---|
| Transaction — Exchange | 93,426 | 95,827 | 56,364 | −3% (declining) |
| Transaction — OTC | 4,013 | 2,370 | 1,531 | +69% |
| Transaction — Other (NFT) | 581 | 879 | 2,617 | −34% (wound down Jan-26) |
| Total transaction | 98,020 | 99,076 | 60,512 | −1% (flat) |
| Services — Credit card | 33,120 | 11,632 | 5,808 | +185% |
| Services — Staking | 16,774 | 11,480 | 823 | +46% |
| Services — Custodial fee | 8,739 | 7,004 | 4,104 | +25% |
| Services — Advisory (warrants) | 4,808 | 0 | 0 | new |
| Services — Other | 1,198 | 0 | 0 | new |
| Total services | 64,639 | 30,116 | 10,735 | +115% |
| Interest income (in net rev) | 11,434 | 11,672 | 25,066 | −2% |
| Total revenue | 179,572 | 142,165 | 98,137 | +26% |
The headline (+26% revenue) flatters a worrying mix shift: the core exchange business is shrinking (−3%, on a 4% retail-volume decline) ``, and total transaction revenue is flat. The +26% is almost entirely services — a 185% credit-card surge off a tiny base and a 46% staking rise. Geographically the business is US-concentrated and about to become more so post the UK/EU/AU exit. Trend = the growth engine (services) is real but small; the value engine (exchange) is decelerating.
Income statement, Q1 2026 vs Q1 2025 ($000s) ``:
| Q1 2026 | Q1 2025 | |
|---|---|---|
| Net revenue | 48,578 | 35,117 |
| Total revenue | 50,272 | 35,322 |
| Total operating expenses | 144,460 | 83,349 |
| — Salaries & comp | 65,428 | 34,272 |
| — Sales & marketing | 19,071 | 9,036 |
| Operating loss | (94,188) | (48,027) |
| Net loss | (108,978) | (149,264) |
| Diluted EPS | $(0.93) | $(30.34) |
| Wtd avg shares (000) | 116,582 | 4,920 |
. But the **operating loss nearly doubled** to $(94.2)M, because opex grew 73% YoY (salaries +91%, S&M +111%, SBC $24.2M vs $1.5M) . The net-loss "improvement" to $(109.0)M is an artifact of FY25's huge related-party crypto-loan fair-value swings rolling off, not operating progress.: Cash $215.6M + restricted $103.7M; **related-party loans still $252.6M**; funding debt $140.5M; accumulated deficit **$2.15B**; total equity $456.1M (down from $540.9M one quarter earlier). **Operating cash burn $(54.4)M for the quarter** — at that run-rate, unrestricted cash covers ~4 quarters absent new financing .— but the swing was driven less by the beat than by a **$100M Winklevoss Capital Fund injection** announced alongside it. The market rewarded a related-party bailout, not earnings power.Unusual vs own history: the post-IPO comp + marketing blow-out (opex 73% higher YoY on flat-ish transaction revenue) is the standout anomaly — a company spending into a downturn while its core line shrinks.
No transcripts/ on the shelf (transcripts=0). GEMI has only been public since Sep 2025, so the public-call history is two prints (Q4-2025/FY and Q1-2026). From coverage , the management arc has shifted hard from **growth-story (IPO roadshow: superapp, $1M-bitcoin bullishness)** to **survival/efficiency**: the dominant Q1-2026 messages were the ~25–30% workforce cut, the UK/EU/AU wind-down ("hard to win in for various reasons"), and an AI-productivity pitch ("AI has expanded engineer productivity by 100x") . Recurring new phrase: path to profitability / efficiency. Dropped: international expansion, NFT (Nifty Gateway killed Jan 2026). The "100x AI" claim is the tone tell — promotional framing layered over a retrenchment (flagged again in Lens 9/13). Label: directional, web-sourced; no transcript on disk.
Crypto-exchange / fintech-brokerage peer set. Multiples are `` with date, or n/a.
| Company | Ticker | Mkt cap | EV/Sales | P/E | Notes |
|---|---|---|---|---|---|
| Gemini | GEMI | ~$0.48–0.50B `` | ~2.5–3x `` | n/a (loss-making) | Negative adj. EBITDA $(258)M FY25 |
| Coinbase | COIN | ~$44.7B `` | ~7.0x `` | ~57x `` (16.8x earlier Apr-26 — volatile) | 8.6% volume share |
| Robinhood | HOOD | ~$83B `` | ~20.8x `` | ~51x `` | Diversified brokerage |
| Circle (stablecoin peer) | CRCL | n/a | n/a | n/a | GUSD comp by mechanism |
5-yr avg ROE: n/a for GEMI (only public since Sep-2025; ROE is deeply negative on a $2.15B accumulated deficit). The honest read: GEMI optically looks "cheap" on EV/Sales (~2.5–3x vs COIN ~7x) only because it is unprofitable and shrinking in its core line — a value trap multiple, not a discount. You do not pay a revenue multiple for a business burning $250M+ of adjusted EBITDA a year; you pay on a path to cash generation, which is not yet visible.
GEMI has <1 year of tape, but the moves are violent and instructive ``:
What the tape reveals: the market reacts to (1) crypto-cycle beta (BTC direction), (2) liquidity/solvency signals — the single biggest up-day was a founder cash injection, the biggest down-leg was the retrenchment — and (3) the profitability question. This is being traded as a distressed, cycle-levered going concern, not a growth compounder.
. They built a genuinely pioneering regulated exchange (first regulated ETH listing, early NYDFS license, GUSD). But the *operating* track record is weaker than the *pioneering* one: a $2.15B accumulated deficit , sub-1% share after 11 years, and an IPO that lost ~85% of its value in nine months.Forensic lens. Every figure labeled.
. The lending agreements **have no covenants and no maturity, and WCF can demand repayment at any time on written notice** — a call option on the company's liquidity held by its controlling shareholders. Fair-value remeasurement of these related-party instruments drove a $243.1M "other expense" swing in FY2025 (incl. a $119.3M change in fair value on related-party loans), the proximate cause of the $582.8M net loss ``. Earnings are being whipsawed by accounting on insider debt., and adj. EBITDA explicitly addbacks SBC as a "significant recurring expense" — non-GAAP flatters a real, ongoing dilution cost.Regulatory findings (required sub-section):
total_sec_findings: 0 — "No LR found" and "No AAER found" for Gemini Space Station 2021–2026 ``.. The substantive litigation is in the risk factors: **as of 1 Mar 2026, 373 arbitrations by Gemini Earn users** seeking additional interest beyond their Genesis-bankruptcy recovery; Gemini Trust agreed to 15 "bellwether" arbitrations (won the first 5), with the rest stayed; plus 49 settled IRAF-client arbitrations . Also a pending DC EFTA jurisdictional appeal ``.. CryptoSlate flagged residual risk that SEC settlement terms could constrain future yield products .Built bottom-up from FY2025 actuals + Q1-2026 run-rate , with explicit assumptions . This is a path to narrower losses, not to EPS — GEMI has no positive EPS in the visible window.
Revenue base: FY2025 total revenue $179.6M; Q1-2026 $50.3M ≈ $201M annualized run-rate ``.
| Scenario | FY2026 revenue | FY2026 op. assumptions | FY2026 net result |
|---|---|---|---|
| Bear | ~$165M `` | Layoff savings lag; S&M sticky | Net loss ~$(280)M; cash runway pressured into 2027 |
| Base | ~$205M `` | Net loss ~$(180)M to $(220)M; adj. EBITDA improves from $(258)M but stays negative | |
| Bull | ~$260M `` | Operating leverage on cost cuts | Net loss narrows to ~$(90)M; line-of-sight to adj-EBITDA breakeven in FY2027 |
FY2027–28: contingent almost entirely on (a) the crypto cycle and (b) whether prediction markets / credit card scale into a second leg. In the bull, adj-EBITDA breakeven is plausible FY2027 and modest GAAP profitability FY2028; in base/bear, losses persist and the question becomes how many more WCF injections the founders fund. The honest base case: still loss-making across the three-year window, survival underwritten by founder capital.
Brier forecast: --watchlist rule says skip forecast.ts create in the loop (no forecast.ts create run — per skill + wave boundary). The natural scoreable claim if promoted: "GEMI FY2026 adjusted EBITDA remains negative," p≈0.90.
Bull case. Gemini is a real, fully-licensed US crypto exchange + the only NYDFS-regulated stablecoin franchise of scale outside Circle, trading at ~$0.5B — a small fraction of Coinbase. If (1) the crypto cycle turns up, (2) the 30% cost cut + US-only focus drives genuine operating leverage, and (3) services (credit card +185%, staking, CFTC-regulated prediction markets vs Kalshi/Polymarket) compound into a credible second leg, then a $0.5B EV on $200M+ of growing, regulated revenue with founders who will fund it is a high-torque cycle call. Secular tailwind: the GENIUS/CLARITY regulatory normalization legitimizes the whole category. Earnings surprise potential is asymmetric off a beaten-down base.
Bear case (2–3 permanent-impairment risks). (1) Subscale is structural, not cyclical — <1% volume share vs Coinbase's 8.6% ``; in an exchange business, liquidity begets liquidity, and Gemini is on the wrong side of that flywheel. The core exchange line is already shrinking (−3% FY25) during a crypto up-leg. (2) The capital structure is a founder call option — un-covenanted WCF loans repayable on demand + recurring rescue injections mean public equity is structurally subordinate to insider liquidity decisions; a founder change-of-heart, or simply tapped-out patience, is an existential risk. (3) Cost base built for a bigger company — even post-layoff, opex ($525M FY25) dwarfs revenue ($180M); the path to profitability requires either a doubling of revenue (cycle-dependent) or further deep cuts (capacity-destroying).
Pre-mortem (18 months out, thesis broke): crypto chopped sideways/down through 2026–27; exchange volume share kept eroding to the majors and to offshore/DEX venues; credit card and predictions never scaled past rounding-error revenue; the company burned through unrestricted cash, took another dilutive WCF round (or a down-round convert), and the stock re-rated to a sub-$300M distressed multiple. The "trusted brand" never converted into pricing power because regulatory clarity commoditized compliance.
Are multiples too high? On EV/Sales (~2.5–3x) it looks cheap vs COIN, but that's a value trap — the right frame is enterprise value vs sustainable free cash flow, which is negative. There is no earnings multiple to be too high; there is a solvency-and-cycle bet.
Contrarian view (what the market refuses to see): Both bull and bear consensus treat GEMI as a pure crypto-cycle option. The thing the market under-weights is that the founders' willingness and ability to backstop is itself the thesis — this trades less like an exchange and more like a perpetually-recapitalized founder vehicle. That caps the downside (they won't let it fail cheaply) and the upside (public holders get diluted on every rescue). The asymmetric bet, if any, is on a cycle-driven volume spike monetizing the fixed cost base before the next raise — narrow, and timing-dependent.
Dismantling the bull case. What structurally breaks how Gemini makes money: it is sub-scale in a winner-take-most business and its core revenue is already in decline at the top of a cycle. ~50% of volume sits with a handful of market makers who can leave; ~55% of revenue is transaction fees that are pure crypto-volume beta with no contractual floor . The most dangerous competitor bulls underrate isn't Coinbase (priced in) — it's the combination of **Robinhood** (mass-retail crypto at a fraction of the friction) and **offshore/DEX venues** that offer products and leverage Gemini legally can't, hollowing out exactly the active-trader cohort Gemini needs. The worst capital-allocation pattern is the **related-party financing loop**: founders lend at 4–8%, collect $24M/yr in fees, then convert to super-voting stock and inject rescue equity — a structure that extracts from and dilutes public holders simultaneously . For today's ~$0.5B price to hold, you must assume both a crypto bull-cycle and successful cost discipline and a services second-leg — three things, all must hit. If revenue disappoints 20–30% (entirely plausible in a crypto drawdown — exchange revenue fell even in 2025), the company is back to burning $250M+ against a sub-$170M top line and dependent on the next WCF cheque. The single permanently-impairing scenario: a multi-quarter crypto bear market that drains unrestricted cash while public-market appetite for dilutive raises is shut — forcing a deeply dilutive insider recap that resets the equity to distressed levels. Plausibility: moderate-to-high over a 2-year horizon, given the cash burn and cycle dependence.
A profitable, fast-growing digital bank mispriced as a crypto name and now mispriced again by fear — but the entire bull/bear hinges on one binary question the market cannot yet resolve: are the +$2.0B of fair-value loan marks real, or is Muddy Waters right?
A levered, sub-NAV Ethereum proxy run by the best operators in the trade — own the discount-closing buyback, not the "treasury premium" that already died; the bet is ETH plus a 0.83x→1.0x mNAV mean-reversion, fighting a 95% drawdown and a structurally bigger rival.
A real, cash-generative super-app now valued like a crypto-beta call option — the bull thesis (tokenization + a 27M-customer flywheel) is genuine, but ~37x forward earnings re-prices violently every time the crypto/options trading cycle exhales, and the prediction-markets engine that drove Q1 growth is one adverse court ruling from being switched off.