Coinbase (COIN): The Right Company at the Wrong Price

Coinbase runs the best regulated franchise in crypto — and at $165 the stock prices a full bull-market recovery as if it were the normal state of the world. A through-cycle valuation of what the business actually earns, what today's price assumes, and what would change our mind.

TON618 Capital Research · COIN $165.48 · BTC $62,632 · as of 4 Jul 2026 · Download PDF

Update, 30 Jul 2026 — Q2 2026 results: addendum — The Scissors Open Wider (trigger check; fair value unchanged).

1Executive Summary

COIN at $165 prices a full cyclical recovery as the base case; on through-cycle normalized earnings we estimate fair value at ≈$102 per share (range $85–$120), roughly 40% below the market price. Even our bull scenario (BTC $110k+, full retail return) supports only ≈$201 — modest upside for a stock this volatile — because today's price already embeds most of that outcome. A corollary that investors weighing COIN as a vehicle for crypto exposure should see plainly: in every scenario we model, COIN delivers a worse return than simply holding Bitcoin directly, since the equity's valuation premium consumes the operating leverage that is its reason for existing. The analysis below sets out the assumptions, the upside case against our estimate, and the specific evidence that would change our view.

The Current State of Coinbase's Story

Coinbase is the largest regulated cryptocurrency exchange in the United States. It makes money in two very different ways. The first is trading: when people buy and sell crypto on its platform, Coinbase takes a cut. This business is enormously profitable when crypto prices are rising and retail enthusiasm is high — and it shrinks brutally when they fall. It is shrinking now: with Bitcoin down about 50% from its late-2025 peak, everyday traders have gone quiet, and the fees Coinbase collects from them have fallen 48% in a year. The second business is quieter and steadier: Coinbase is becoming crypto's financial plumbing. It safeguards the coins inside most US Bitcoin ETFs, it runs the world's largest crypto options venue (Deribit, acquired in 2025), and it collects roughly half the interest earned on the reserves behind USDC, the second-largest dollar-pegged digital currency (a "stablecoin"). That second business kept growing right through the downturn.

Data shows the company is financially sound: more cash than debt, real profits in good years, and roughly breakeven operations even in the current trough. The question is not survival — it is price. We believe the stock, at $165, is priced as if the good years are the normal years. Our estimate of fair value — built by weighing a bad outcome, a middling one, and a strong recovery by their likelihoods — is about $102 per share. To justify today's price, you must assume crypto activity returns to its 2025 peak levels and stays there, an outcome we think deserves about a one-in-four weight, not a base case.

The strongest argument against our caution deserves a fair hearing: the plumbing businesses — stablecoin interest, custody, derivatives — are growing, high-quality, and increasingly independent of the trading cycle. If investors ever value them the way they value software companies, our framework is too conservative and the stock is worth considerably more. That is a real possibility; we weigh it explicitly in the risks section. But at ≈$2.4 billion of yearly revenue against a $43 billion valuation, the plumbing alone cannot yet carry the price. One more observation for anyone buying the stock as a stand-in for Bitcoin itself: at this price, our scenarios all show the stock delivering less than the coin — the premium you pay for the company consumes the extra upside its business would otherwise provide.

2Investment Context

• Cycle position: BTC peaked ≈$126k in October 2025 and sits at $62.6k, −50%, nine months in. Historical crypto bears have bottomed 12–18 months from the peak at −77% to −85%, implying a bottom window of late 2026 with recovery earnings arriving 2027–28, not immediately.
• The stock has already been repriced for the winter: COIN is −63% from its 52-week high ($444.65), with 50/200-day moving averages at $179/$231.
• What the security is: COIN is a high-beta (raw beta 3.35) equity claim on crypto activity (trading volumes × take rate, plus a growing services layer), not on crypto price. The valuation question is therefore not only "is COIN cheap in absolute terms" but "what does it offer relative to direct crypto exposure" — we address both.
• Structural changes since the last cycle: S&P 500 inclusion (May 2025); the Deribit acquisition closed August 2025 ($4.3B of consideration at closing — announced at ≈$2.9B, the difference being COIN share appreciation between signing and close; the world's largest crypto options venue — now driving institutional transaction revenue +37% YoY even in the bear); federal stablecoin regulation (GENIUS Act, 2025) supporting USDC; the Base L2 platform.

3Business & Revenue Architecture

Revenue disaggregation (SEC EDGAR, 10-K FY2025 / 10-Q Q1'26):

FY2024 rows sum to $6,298M; the ≈$266M balance (principally interest and finance-fee income) was not disaggregated into these buckets in that filing vintage — it is inside "custody + interest + other" in the FY2025 and Q1'26 columns.

Quarterly path: 2,034 → 1,497 → 1,869 → 1,781 → 1,413 (Q1'25→Q1'26). (Q4'25 derived from FY minus Q1–Q3 per EDGAR; Alpha Vantage's $1,032M Q4 figure is wrong and was discarded.)

The two-business framing. Coinbase is (a) a hyper-cyclical retail trading franchise with ≈70%+ incremental margins, stapled to (b) a growing, sticky "crypto financial infrastructure" annuity (stablecoin + custody + derivatives + Base) now running ≈$2.4B/yr and growing through the bear. Business (b) is the bull's best argument; it is why COIN deserves a premium to exchange multiples on normalized earnings. It is not, at ≈34% of trough revenue, large enough to support a $43B franchise valuation on its own.

4Financial Statement Analysis & Quality of Earnings

We screened the financial statements against a standard forensic checklist (Beneish/Schilit-style earnings-quality flags) before valuing. Escalation to a full forensic work-up is not warranted — the findings, transparently:

• Cash conversion is strong: FY25 operating cash flow $2,426M vs net income $1,260M (1.9×); capex ≈ 0. No accrual buildup, no receivable/inventory games (no inventory; receivables stable).
• Fortress balance sheet: $10.2B cash + $0.5B securities vs $7.2B of senior/convertible notes plus $0.2B finance leases (incl. $2.96B new converts issued Q4'25) → net corporate cash ≈$3.3B (a further $0.6B of short-term collateralized borrowings sits inside the crypto-lending book, offset by collateral held), plus a $1.6B (book, 3/31) crypto investment portfolio: 16,492 BTC, 150,193 ETH, $163M other. Customer custodial funds ($5.5B) are properly segregated and offset by matching liabilities.
• The Q1'26 loss is mark-to-market, not operational: the −$394M net loss was driven by −$482M of non-operating crypto marks; operating income was ≈breakeven (−$21M) on a 31% revenue decline — evidence of cost discipline learned in 2022–23 (a restructuring was taken in Q1'26).
• Flags worth carrying (disclosed, not disqualifying): (1) $4.2B total goodwill ($2.8B of it from Deribit) + $1.4B intangibles, booked at the cycle peak — impairment risk if derivatives volumes fade (non-cash but a headline risk); (2) stock-based compensation of $839M = 12% of revenue → persistent ≈3–4%/yr dilution (we use 288M diluted shares); (3) a $1.4B loan book collateralized by $1.1B of crypto collateral held — overcollateralized but pro-cyclical credit exposure; (4) FY24's $2.58B net income included ≈$0.6B of non-operating gains — readers should not anchor on it.
• Auditor Deloitte; timely filings; no restatements. The earnings problem at COIN is cyclicality, not accounting.

DuPont view: FY25 ROE ≈8.9% (NI 1,260 / avg equity ≈14.1B) — modest, because equity is bloated by cash and the crypto portfolio; the franchise itself is capital-light with high incremental margins.

5Valuation

5.1 Cost structure calibration

EBIT ≈ revenue − 22% × (transaction + staking revenue) − fixed opex. This calibrates to FY25 actual EBIT within $5M (model 1,440 vs 1,435) and to Q1'26 annualized ≈breakeven. Current fixed base ≈$4.9B (post-Deribit); FY25 was ≈$4.7B.

5.2 Scenario valuation (normalized FY2027, $M except per share)

Balance-sheet layer: net corporate cash $3.33B + investments marked to spot (BTC $1.03B, ETH $0.26B, other crypto and strategic holdings incl. the Circle stake at a 15% haircut, $0.84B) = $5.5B, $19/share.

Probability-weighted fair value ≈ $102. Sensitivity on the base case: even at 32× normalized NI +30%, the model tops out at ≈$125. To defend $165 fundamentally, one must treat our bull scenario as the base case — a coherent position, but it should be held knowingly (we weight it 25%; the market implicitly weights it ≈55%+).

5.3 Reverse valuation — what $165.48 requires

Market franchise value (market cap less cash & portfolio) ≈ $42.2B, requiring normalized net income of $1.69B at 25× ($5.86 EPS) or $2.11B at 20×. For reference, FY25 — a good year with BTC averaging ≈$100k — produced $1.26B of net income ($4.45 EPS), and the trailing twelve months produced ≈$0.8B ($2.78). The price embeds FY25-or-better conditions as the permanent norm, plus growth.

5.4 Relative valuation cross-check

COIN's ≈5.9× franchise EV/sales carries a through-cycle EBIT margin of only ≈15–20% (vs CME ≈60% at ≈13× sales) — margin-adjusted, COIN is the most expensive name in the set except HOOD. Street consensus (median price target $236; FY27 EPS ≈$5.08 on $7.6B revenue) effectively assumes FY25 conditions return by 2027 and applies a >30× multiple; we probability-weight that outcome at 25%. (Third-party consensus feeds for COIN are internally inconsistent and show wide cross-vendor dispersion — aggregator price-target medians range roughly $227–$307 depending on source and date. The EBITDA lines were not relied on; the price-target and EPS figures are indicative only.)

A note on how we use comparables. Relative multiples are a pricing tool: they measure what the market is willing to pay today for similar businesses, and we use them for exactly that — they set the 25–28× scenario multiples (deliberately above every mature-exchange comp) and they power the reverse valuation in §5.3. They do not carry the conclusion, for one reason: crypto-linked equities re-rate as a bloc, so comping COIN against its peer set imports the sector's current sentiment as if it were information — relative valuation is least reliable at cycle extremes, which is precisely when a fair-value estimate matters. The practical consequence is that this note's disagreement with the market price is a disagreement about earnings (what a normal year produces), not about multiples (what the market pays for it) — a formulation we prefer because earnings disagreements are settled by results on a schedule, while multiple disagreements can persist indefinitely.

6COIN Versus Direct Bitcoin Ownership

Many investors buy COIN as a levered proxy for the crypto cycle, so a fair valuation should test that use directly. Comparing model outcomes to spot BTC from today's levels:

Because the equity already prices the bull scenario, COIN underperforms direct Bitcoin ownership in every state we model — a 3.3-beta instrument whose valuation premium consumes its torque. On these numbers, the proxy use-case only re-emerges below ≈$120, where base-case outcomes for the stock and the coin converge. This is a statement about price, not about the company: the same franchise at a lower price is a different proposition.

7Special Situation: A Coinbase–Circle Combination

Primary sources: Circle FY2025 10-K (acc. 0001876042-26-000062, filed 2026-03-09); CRCL $64.62, market cap $17.3B (≈267M shares). No direct prediction market exists on this event (Polymarket checked 2026-07-04), so probability is estimated with a conditional-gate framework and disclosed gate by gate.

7.1 Why the scenario is live

1. Revealed interest, both directions. Circle explored a sale before its June 2025 IPO — Ripple's reported $4–5B bid was rejected, and Coinbase was widely reported as the natural acquirer. Circle chose the IPO at what proved to be the cycle top (CRCL peaked ≈$263 on a closing basis, ≈$299 intraday; now $64.62, −75% from the closing peak).
2. The price has collapsed into feasible range. At $17.3B, a 35% premium implies ≈$23B — large but no longer absurd against COIN's $44–48B capitalization. At Circle's $60B+ peak valuation the question was moot.
3. A hard catalyst is weeks away. The Collaboration Agreement (August 2023) has an initial three-year term lapsing ≈August 2026, with a good-faith renegotiation discussion required before automatic three-year renewal. Renewal negotiations and acquisition negotiations are the same conversation held at different stakes.
4. Coinbase holds unusual contractual leverage. Circle's own risk factors disclose that the Collaboration and IP License Agreements can force assignment of Circle trademarks — including the USDC brand — to Coinbase upon certain trigger events. Coinbase negotiates any deal from a position of embedded control.
5. GENIUS Act stakes. With a federal payment-stablecoin framework in force and banks/PayPal entering, owning the #2 global stablecoin issuer outright is a strategic asset — and, defensively, prevents a third party (a bank consortium, Ripple) from buying Circle and squeezing Coinbase at the next renewal.

7.2 Deal math (illustrative, current terms)

• Consideration: ≈$23.3B at a 35% premium ($87/CRCL share). COIN's net corporate cash is $3.3B, so this is ≈85% stock: ≈$19–20B of COIN at $165 = ≈120M new shares, +41% dilution (288M → ≈407M diluted). Coinbase's existing small CRCL stake (inside the $832M strategic line) trims this only modestly.
• What is acquired, economically: Circle FY2025: reserve income $2,637M (+59% YoY), total revenue $2,747M, distribution costs $1,662M — of which ≈$1.35B is the intercompany payment to Coinbase that consolidates away. Circle's FY25 net loss (−$70M) is an artifact of a $424M stock-compensation charge triggered by the IPO (per Circle's 10-K); adding it back yields ≈$320M of normalized pre-tax income, or roughly $230–250M of normalized standalone net income at a normal tax rate (FY24 actual: $157M from continuing operations, on a much smaller reserve base). Add ≈$100M of plausible cost synergies → ≈$320M of incremental normalized net income to the combined entity at current rates and ≈$60B average USDC supply.
• Accretion test: fails at current scale. ≈$23B for ≈$320M incremental NI is ≈70× — paid for with COIN shares this model values at 25× normalized earnings. On our base case the deal is meaningfully dilutive (≈$74 pro-forma fair value per share vs $83 standalone). It turns accretive only if USDC supply roughly doubles (≈$120B) with reserve yields ≥4%, lifting the acquired stream toward $1.1–1.3B NI (≈18–20× paid) — i.e., the acquisition is a leveraged bet on the stablecoin bull case, the same bet the market is already making in COIN's multiple.
• The counter-argument that matters most: Coinbase already extracts ≈81% of Circle's distribution line (≈51% of the entire reserve-income pool) by contract, without owning the issuer's regulatory capital burden — and its 2023 decision (dissolve Centre, take stake-plus-contract instead of buying) reveals that preference. Acquisition buys the residual ≈$300–400M stream, brand control, and renewal-risk elimination; it is defense, not offense.

7.3 Would buying Circle be good for Coinbase? The bull case, the bear case, and the common-sense view

The bull case. Coinbase would own crypto's dollar rails outright. USDC is a machine that converts short-term interest rates into revenue at almost no marginal cost — every dollar in circulation is backed by Treasury bills earning yield — and today Coinbase collects only its contractual share of that machine, under an agreement that comes up for discussion every three years. Buy Circle, and the renewal anxiety disappears forever; the other half of the income pool comes in-house; and Coinbase becomes the complete regulated stack — the exchange, the ETF custodian, the largest derivatives venue, and the stablecoin itself — at precisely the moment federal law has turned stablecoins into a legitimate payments industry. The bull adds a defensive kicker: buying Circle keeps anyone else (a bank consortium, Ripple) from owning Coinbase's most important business partner and squeezing it at the next renewal. And the timing argument writes itself: the target is available at a 75% discount to where it traded a year ago. If stablecoins end up embedded in mainstream payments — the bull's end-state — the acquired income stream triples, and today's price will look like a footnote.

The bear case. Coinbase would pay roughly $23 billion — half its own market value, mostly in shares this note argues are themselves richly priced — for roughly $320 million of additional normalized profit. That is ≈70× earnings for a business whose revenue Coinbase already mostly receives by contract, without bearing the issuer's burdens. Ownership brings those burdens home: regulatory capital requirements, bank-style supervision, and the tail risk that defines every stablecoin — if USDC ever "breaks the buck," the issuer stands in front of the run. The deal would also double down on interest-rate exposure at the top of the rate cycle: Fed cuts compress the entire acquired stream mechanically. And the worst version of the story is strategic, not financial: paying a control premium for the incumbent stablecoin at the exact moment federal legislation invites banks — with cheaper funding, bigger balance sheets, and existing payment networks — to commoditize the product. In that world Coinbase would have bought the category's Kodak.

The common-sense view. Strip away the finance and the situation is homely: Coinbase is a shopkeeper who already collects most of the rent from a tenant's thriving stall, under a lease that even lets the shopkeeper seize the tenant's brand name if things go wrong. Buying the whole building ends the lease anxiety — but it costs half of everything the shopkeeper owns, and it makes him the landlord of the risks as well as the rents. Whether that trade is wise depends entirely on price and on which future arrives. Coinbase has faced exactly this decision once before, in 2023, and chose the contract over the company; nothing has changed since except that the company got cheaper — and cheapness alone is not a strategy. The sensible path, and probably the likely one, is to use the August 2026 renewal to extract better terms while holding the acquisition in reserve as a defensive weapon, to be drawn only if someone else reaches for Circle first. Our verdict, plainly: at ≈$23 billion the deal is better for Circle's shareholders than for Coinbase's; below ≈$15 billion, or in a world where stablecoin adoption is visibly doubling, the answer flips.

7.4 Obstacles

• Governance: Circle's multi-class structure concentrates voting power with founder Jeremy Allaire, who chose independence at the IPO one year ago. A board that listed at $31 and watched the stock trade to $263 will anchor well above $87.
• Regulatory: HSR review of a vertical stack (largest US exchange + ETF custodian + #2 stablecoin) is a real but likely surmountable hurdle given Tether's dominance and bank entrants; Circle's pending OCC national trust charter adds banking-regulator sign-off complexity and timeline.
• Currency: paying a control premium in COIN shares that are themselves cyclically depressed (in our estimate ≈65% above fair value, though management would not agree) transfers less value than it appears — this cuts for a stock deal from Coinbase's perspective and against Circle accepting one.

7.5 Probability estimate

Conditional-gate build, 24-month horizon:

Best estimate: ≈12% (range 10–15%) that Coinbase acquires Circle within two years. The single best tell will be the ≈August 2026 renewal outcome: a quiet three-year auto-renewal on unchanged terms kills most of the near-term probability; a renegotiated split, an announced "expanded partnership," or a renewal delay marks the scenario live.

7.6 Impact on the valuation

Probability-weighted, the event is roughly valuation-neutral for COIN: ≈12% × (−$9/share base-case dilution, +$20–30/share in the stablecoin-bull tail) ≈ ±$2–3/share. Fair value stays ≈$102. The larger point is informational: the conditions under which Coinbase would buy Circle (stablecoin scale-up, sustained rates) are the same conditions under which our bull scenario pays — the deal is a symptom of the bull case, not an independent source of upside. For CRCL holders the asymmetry is different (a 35% premium at 12% probability is worth ≈4% of expected value on top of standalone value), but CRCL is outside this note's scope.

8Risk Factors (to this view — i.e., what makes our fair-value estimate too low)

1. Stablecoin/infrastructure re-rating (highest likelihood, high impact): USDC income (+11% YoY in a bear) plus Deribit could get valued as a standalone SaaS-like annuity at 10×+ sales, supporting $60–80/share for the infrastructure businesses alone; a decisive GENIUS-Act-driven payments breakout would break our multiple framework upward.
2. A faster cycle: If BTC bottoms early and 2027 is a full bull year (consensus's implicit view), FY27 EPS of $5–8 at 30× supports $170–250. We assign this 25% probability; the market assigns materially more.
3. Rate path: Stablecoin revenue is reserve-yield-sensitive; slower Fed cuts mean upside to our stablecoin line (and vice versa — cuts to 3% could cost ≈$300M of revenue on flat supply).
4. Downside risks (what makes our estimate too high): Deribit goodwill impairment; loan-book stress in a deeper crypto drawdown; fee compression on consumer take rates (the long-running secular threat); regulatory reversal under a different administration.

• Circle revenue-share renegotiation: the stablecoin line depends on the Coinbase–Circle distribution agreement (≈100% of reserve income on platform-held USDC, ≈50% of residual off-platform). Distribution cost to Coinbase is Circle's largest expense and its clearest margin lever; Circle has been diversifying distribution (e.g., Binance, late 2024), and the agreement's initial term lapses ≈August 2026 (see Section 7). Our scenarios assume current terms persist.

1. Model risks: normalized-multiple selection (20–28×) is judgment; a strict CAPM at COIN's raw beta of 3.35 (cost of equity ≈ 23%) would value the franchise far lower still — our multiples are already generous against an FCFE perpetuity (25× NI implies Re−g ≈ 4%).
2. Framework limitation — mean reversion vs. a compounding platform (the most important upside risk to our estimate). Our normalization treats COIN as a cyclical oscillating around a stable mid-cycle level. Coinbase's own history argues partly against this: bear-market revenue floors have ratcheted — the 2023 trough year produced $3.1B of revenue, while the current trough run-rate is ≈$5.65B (+82% trough-over-trough). If that ratchet persists, our "normalized" base is really the floor of the next cycle. Accepting the compounding-platform view in full — our bull revenue path as the mid-cycle norm at 28× — supports ≈$160–200, which brackets the current price: the market disagreement with this note is about which model of the company is true, not about the arithmetic. We hold the mean-reversion frame because Coinbase's premier position sits in a genuinely competitive industry (retail take rates ≈10× offshore peers, under attack from zero-fee entrants), unlike the monopoly-structure franchises where compounding-platform valuations have paid off — and because the identical premier-name argument was made for COIN at $357 in 2021, eleven months before it traded at $31. The discriminating evidence: if stablecoin/custody/derivatives keep growing through this winter and the trough revenue floor holds above ≈$5.5B, the compounding view gains weight and fair value migrates toward $130–160 without any bull market required. We will update on that evidence.

9Catalyst Calendar

Dated events over the next 12–18 months, with direction and estimated magnitude against our ≈$102 fair value. The dominant driver — the crypto cycle itself — is deliberately listed last: it has no date, and it outweighs everything above it.

Reading the table honestly: every dated catalyst is worth single-digit to low-double-digit dollars per share; the undated one — where Bitcoin is in its cycle — spans $42 to $201 across our scenarios. Investors positioning around the dated events are trading the noise of a name whose signal is the cycle.

10Ten Open Questions, Answered Against the Model

The questions investors are actually asking on COIN right now, answered with the verified data and the framework above. Where a premise in circulation conflicts with the primary filings, we say so.

1. What would CLARITY Act passage actually mean for COIN — and does it pass? The state of play first: the bill cleared Senate Banking 15–9 in May and has sat on the Senate Legislative Calendar since June 1 — floor-vote-eligible at any time, with a pre-July-4 vote postponed over a conflict-of-interest provision and law-enforcement objections to §604. Markets price full enactment at ≈6–7% before August 1, 2026 (Kalshi) and ≈41–48% during 2026 (Kalshi/Polymarket, the latter down from 74% a month ago); no scenario in this note assumes enactment, so passage is unpaid-for upside. Because the floor vote can be scheduled any week, this is the nearest-dated meaningful catalyst on the calendar (§9) despite the low by-August odds. The more important question is transmission — what passage is worth:

• Four channels of value. (i) Tail-risk removal: a statutory commodity/security classification ends the reclassification threat hanging over the listable universe and the $677M staking line — this shows up as multiple support, worth roughly 2–3 turns on normalized earnings (+$5–8/share). (ii) Listing expansion: tokens exit securities limbo and become listable — modest, incremental trading revenue. (iii) The real prize — tokenized securities: a federal registration path toward trading digital-asset securities is what converts the "Everything Exchange" from pitch to permission, opening a TAM measured against securities-market infrastructure rather than crypto trading. This is genuine optionality, but it monetizes over years, not quarters. (iv) Institutional unlock: banks and asset managers get a rulebook to allocate against — supportive of custody and institutional volumes.
• Putting numbers on it: mechanically, passage justifies lifting the base multiple toward the top of our band and shifting scenario weights bull-ward — together worth roughly +$15–30 per share of fair value (≈$102 → ≈$120–130). Meaningful, but note the proportions: the legislation is worth perhaps 20% on fair value while the crypto cycle spans $42–$201 in our scenarios. The cycle is roughly five times the size of the law.
• The empirical caution comes from the last law. COIN's all-time high — $444.65 — was set on July 18, 2025, the day the GENIUS Act was signed. The stock has fallen 63% since, because the cycle turned and earnings, not statutes, set prices. Legislation moves the multiple; the cycle moves the earnings; the earnings are bigger.
• The moat paradox, stated honestly. Part of Coinbase's premium is the regulatory murk: it earns a scarcity rent for being the compliant venue in an unclear regime. CLARITY hands every bank and incumbent exchange the same rulebook. The near-term re-rating is real; the long-run effect on take rates and custody share is plausibly negative as compliance stops being a moat and becomes a commodity. Passage is unambiguously good for the industry; for Coinbase specifically it is front-loaded good news with a competitive bill that arrives later.

2. Can subscription & services growth outrun trading-fee cyclicality? The mix shift is real but too slow to de-cyclicalize this cycle. Verified: S&S revenue was $2.83B in FY25 and runs ≈$2.3B annualized at the Q1'26 trough — against a ≈$4.9B fixed cost base, it covers roughly half of fixed costs; consumer trading still decides whether the company makes money. S&S did rise from ≈35% of revenue in FY24 to ≈41% at the trough, but largely because the denominator collapsed. For services alone to carry the cost base it must roughly double — at the stablecoin line's through-bear growth rate (+11%) plus Deribit and custody, that is a 2029–2031 outcome, not 2027. The falsifiable marker is the one in Risk #6: a trough revenue floor holding above ≈$5.5B.

3. Is the Q1'26 stabilization sustainable into Q2? First, two premise corrections: Q4'25 revenue of $1,781M was down 4.7% quarter-over-quarter (from Q3'25's $1,869M), not 31% — the −31% figure is Q1'26's year-over-year decline; and the −$667M Q4 GAAP loss was indeed dominated by unrealized crypto marks, as was Q1'26's −$394M. Q1'26 printed $1,413M of revenue with roughly breakeven operations. For Q2 (reporting expected around July 31 — the company filed its Q2 10-Q on July 31 in each of the last two years): Bitcoin averaged materially lower in April–June than in Q1, so consumer transaction revenue likely fell again, and we would expect a print below Q1's revenue with operations near or slightly below breakeven. That outcome is already inside our trough assumptions; the number that matters in the release is the subscription & services floor, not the trading line.

4. How far and how fast can the "Everything Exchange" diversification go? This is the compounding-platform question (Risk #6) in product form. What is proven: Deribit-driven institutional revenue growing +37% through a bear; the custody moat; Base. What is unproven: equities and prediction-market offerings launching into incumbents — note that Robinhood at 55× earnings is priced as the winner of the same convergence race run from the other direction, so the market is currently paying both companies for the same future. Our bull scenario ($9.15B revenue) embeds meaningful success; the base case deliberately does not pay for unlaunched products. The fastest verifiable markers: the institutional transaction run-rate (now ≈$543M annualized) and any first disclosure of non-crypto trading revenue as a separate line.

5. Do USDC economics scale as projected? Verified baseline: USDC ended 2025 at ≈$75.3B in circulation (average ≈$60B through the year), generating a $2.64B reserve-income pool of which Coinbase's contractual share was $1,349M (≈51%). Treat $1.2T-by-2028 projections as promotional scenarios, not planning inputs: that is ≈16× growth in under three years, and at 4% yields it implies a ≈$48B annual pool whose Coinbase half would exceed the company's entire current revenue — a useful reductio for why such projections should not be capitalized. Our base case carries stablecoin revenue to $1.45B by FY27. The genuine open items are the ones in Section 7: renewal terms at the ≈August 2026 agreement lapse, the yield-bearing-stablecoin rulemaking fight with the bank lobby, and the rate path (each 100bp of Fed cuts costs the pool roughly a quarter of its income on flat supply).

6. How tethered does COIN stay to Bitcoin sentiment? The market has already voted: beta of 3.35, and the stock is down 63% against Bitcoin's 50% — COIN currently trades with a negative decoupling premium (it amplifies the cycle rather than dampening it). On the current revenue mix we think that is correct, and Section 6 quantifies the consequence. A decoupling premium has to be earned in the data, not narrated: the evidence that would earn it is two or three consecutive quarters of services growth against a falling or flat Bitcoin price, at which point the correlation regime — and our multiple — would both deserve revisiting.

7. Which price-target regime is right — the Street's $250+ or bearish models near $150–185? Our feeds do not corroborate a $345 consensus: FMP shows consensus $236 / median $230 (range $107–440), and cross-vendor aggregator medians run roughly $227–307. The dispersion itself — across vendors, not just analysts — is the signal. Note honestly where this report sits: our probability-weighted $102 is below even the bearish models cited, and the reverse valuation in §5.3 explains the entire gap — targets of $230+ require treating FY25-or-better earnings as the permanent norm at 30×+ multiples. The deeper answer is that point targets communicate false precision on this name: plausible outcomes genuinely span $42–$201 (§5.2), and any single number without its scenario weights is marketing, ours included.

8. Do SEC/digital-securities rulings keep going Coinbase's way? The skew has inverted. Most of the big regulatory wins are already delivered and in the price — the dropped SEC enforcement action, the GENIUS Act, S&P 500 inclusion, ETF custody legitimacy. What remains on the docket (CLARITY at ≈41% by year-end, token-classification rulings) offers incremental multiple support if favorable, while an adverse classification ruling or a policy reversal under a future administration is a larger per-event move against a $42B franchise value. From this price, regulatory news is asymmetric to the downside for the first time in two years.

9. Can margins hold as "adjusted EBITDA" scales? Handle adjusted EBITDA with tongs: Coinbase's definition adds back stock-based compensation, which at $839M is 12% of revenue — Q1'26 printed roughly breakeven GAAP operations while adjusted EBITDA would show a healthy positive number, and the gap is the SBC. This note works on GAAP operating income and a fixed-cost base for exactly this reason. The execution question is real: the fixed base grew from ≈$4.7B to ≈$4.9B absorbing Deribit, and the Q1'26 restructuring shows the discipline muscle still works. The watch-item is annualized fixed opex against our $4.7–5.0B modeling band — if it breaks above that while revenue sits at trough, the bear scenario's economics arrive early.

10. Does the ETF custody moat hold against bank entrants? Keep the economics and the strategy separate. Economically the custody line is small — ≈$140M of custodial fees in FY24, an estimated ≈$250M in FY25 — and even zero growth there costs only ≈$2–3 per share in our framework. Strategically, erosion has already started at the margin: BlackRock added Anchorage as an additional custodian in 2025, several issuers added BitGo, and bank trust charters (BNY, Citi, Fidelity's self-custody precedent) point to fee compression more than share loss near-term, since switching the custodian of an operating ETF is operationally painful. Our valuation does not lean on custody revenue; its real function is as evidence for the infrastructure-quality argument that sets the 25–28× multiple — and that argument survives partial share loss.

11Conclusion

• Fair value: ≈$102 per share (range $85–$120) against a market price of $165.48. The valuation gap is not subtle: the market franchise value requires cycle-peak earnings to be the permanent norm, and the stock underperforms direct Bitcoin ownership in every scenario we model at this entry price.
• What would move our estimate up: stablecoin revenue accelerating past a ≈$1.6B run-rate (re-rating the infrastructure annuity); a decisive early turn in the crypto cycle (raising the bull-scenario weight above 25%); enactment of market-structure legislation (worth ≈+$15–30 per share via multiple and scenario-weight effects — §10, Q1); a value-accretive resolution of the August 2026 Circle renewal.
• What would move it down: a renegotiated Circle split; Deribit impairment; consumer take-rate compression.
• Horizon for the view: 12–18 months, spanning the expected cycle trough. Below ≈$120 the risk/reward changes character; readers should treat that level, not our point estimate, as the actionable boundary of this analysis.