Coinbase Global

Phase A — Understand the business

Lens 1 · Company Overview

Coinbase is the largest US-regulated crypto platform — effectively the compliant front-door between the dollar economy and the onchain economy. Three customer franchises:

1. Consumer — retail's "primary financial account for the onchain economy" (buy/sell/hold/stake, the Coinbase One subscription, the Coinbase One Card).
2. Institutional — a full-service prime brokerage (trading, custody via Coinbase Custody Trust Company, financing/lending, and now derivatives via Deribit).
3. Developers — Base (Coinbase's Ethereum L2), Coinbase Developer Platform, and the USDC stablecoin rails.

How it actually makes money. Revenue splits into two engines:

• Transaction revenue — FY2025 $4,055.4M (Consumer $3,322.8M / Institutional $479.7M / Other $252.9M). This is take-rate on trading volume; it is the cyclical, volume-levered engine. Consumer take-rates are far richer than institutional (Consumer earns ~7x Institutional revenue on a fraction of the volume).
• Subscription & services revenue — FY2025 $2,828.0M (Stablecoin $1,348.8M / Blockchain rewards $677.4M / Interest & finance fee income $247.0M / Other S&S $554.8M). This is the recurring, "diversification" engine management is leaning on.
• Plus Other revenue (corporate interest/income) $297.9M → Total revenue FY2025 $7,181.3M (FY2024 $6,564.0M, FY2023 $3,108.4M).

Key payment-term structure. The single most important contract is the USDC arrangement with Circle: Coinbase keeps 100% of the interest earned on USDC held on its own platform and ~50% of the interest on off-platform USDC. This is a financial-style, interest-rate-sensitive revenue stream, not a product margin — and it is concentrated: one counterparty (Circle) accounted for 19% of total revenue in FY2025, up from 14% in FY2024, rising to 23% in Q1 2026. Customers/competitors/suppliers are largely the crypto networks themselves; the "supplier" relationship that matters financially is Circle (see Lens 2).

Lens 2 · Supply Chain

Coinbase is a financial-infrastructure business, so the "supply chain" is a value chain of liquidity, settlement, and trust. Named stakeholders along the chain:

• Upstream inputs / "raw material":
• Public blockchains — Bitcoin, Ethereum (and the long tail of listed assets). Coinbase does not control these; downtime, congestion, or a chain failure is an unhedgeable input risk it explicitly flags.
• Circle Internet Financial — issuer of USDC; the economic counterparty on the single largest S&S line. Coinbase held ~22% of total USDC supply in early 2025 (up from ~5% in 2022), and Circle paid Coinbase ~$908M of its ~$1.01B total 2024 distribution costs. This is the chokepoint dependency.
• Banking & custody partners / payment processors — fiat on/off-ramps; cash held "at venues" was $110.8M at YE2025. CB Inc. (money transmitter) and CCTC (NY-chartered qualified custodian) sit under NYDFS net-capital rules.
• The company (the platform): matching engine, custody, compliance/KYC, the Deribit derivatives book, Base sequencer.
• Downstream / end customers: 9.2M average MTUs (retail) + institutions (the four counterparties each >10% of the loan book) + developers building on Base.

Chokepoints / single-source dependencies:

1. Circle/USDC — single largest, single-counterparty, and structurally at risk from the GENIUS Act (Lens 3/12).
2. Third-party blockchains — Coinbase cannot control Ethereum/Bitcoin uptime; Base depends on Ethereum L1 for settlement.
3. Banking access — a perennial fragility for any crypto-native; mitigated materially by the OCC trust-charter conditional approval (April 2026, Lens 10).

Lens 3 · Competitive Advantages (moats)

Where the moat is real:

• Regulatory & trust moat (the core). Coinbase is the most-licensed, most-audited, US-public crypto venue — a Deloitte-audited 10-K filer, S&P 500 constituent (added May 19, 2025), and now holder of an OCC conditional national bank trust charter. After FTX, "the regulated one that won't lose your coins" is a genuine switching-cost and brand moat — hard to replicate, and it strengthens as US rules (CLARITY, GENIUS) formalize. This is the durable edge.
• Scale & liquidity in custody. $376.1B of customer crypto held with full keys at YE2025 — deep custody and institutional prime relationships create stickiness that a new entrant cannot bootstrap.
• Base / onchain distribution. Owning an L2 plus the developer platform is a nascent network-effect moat — if onchain activity migrates to Base, Coinbase monetizes the rails, not just the trades.
• Deribit = bought the #1 crypto-options book (~$60B open interest, >$1T traded/yr) — instant scale in the higher-margin, stickier derivatives layer where order-book liquidity is itself the moat.

Where the moat is thin:

• Spot take-rate has no moat. Retail spot fees are the profit engine and the most competed surface — Robinhood, Kraken, Binance, and zero-fee challengers all compress it. Coinbase's blended retail take-rate is a multiple of discount rivals; that premium is a perpetual target.
• Bargaining power is asymmetric the wrong way on USDC — Circle, not Coinbase, issues the token; the GENIUS Act can rewrite who keeps the yield.

Net: a strong regulatory/trust/custody moat around a low-moat, hyper-cyclical core fee stream. That is the whole investment tension.

Lens 4 · Segments

Coinbase reports as a single operating segment (the CODM — co-CEO Armstrong + President/COO Choi jointly — reviews net income and operating income for the company as a whole). So "segments" = the revenue-by-type and geography disaggregation:

By revenue type (FY2025 vs FY2024, $M):

By geography (FY2025): US $6,010.6M / International $1,170.7M — 84% US-concentrated; international is only 16% (no single country >10%). For a company whose growth narrative is "global Everything Exchange," the revenue base is overwhelmingly domestic.

The trend that matters: transaction revenue (the high-quality engine) is structurally flat-to-declining; the entire FY2025 net-revenue growth came from S&S — and within S&S, ~half is the Circle/USDC line that the GENIUS Act now threatens. The mix is shifting from richer, owned trading fees to thinner, counterparty-dependent, rate-sensitive yield.

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print: Q1 2026, the headline event)

Coinbase printed a GAAP net LOSS in Q1 2026 — its second consecutive quarterly loss — and missed hard.:

• Vs consensus: Street expected roughly +$0.27 EPS; actual $(1.49) — a large miss.
• What drove it: (1) crypto trading volumes collapsed — spot volumes −37% QoQ, crypto market volumes −28% QoQ — gutting the Consumer transaction line (−48% YoY); (2) opex rose 8% YoY (Tech & dev $525.6M, up from $355.4M, largely Deribit integration + headcount) even as revenue fell, pushing the operating line itself negative; (3) below the line, a $482.4M mark-to-market loss on Coinbase's own balance-sheet crypto treasury turned a small operating loss into a $394M net loss.
• Margin moves: operating margin went from +35% (Q1'25) to roughly −1.5% (Q1'26). The fall is almost entirely volume deleverage on a fixed-ish cost base.
• Balance-sheet flags: operating cash flow fell to $182.7M (Q1'25 $852.7M, −$670M). Cash & equivalents fell to $10,205.0M (from $11,285.5M at YE2025) — partly because the company bought back ~6.3M shares for ~$1.1B in Q1 2026 (buying its own stock into a loss quarter).
• Market reaction: shares fell only ~4% after-hours — below COIN's historical post-print average, signalling the miss was partly pre-discounted (layoffs had been pre-announced; crypto was visibly soft).

Bridge to the FY2025 10-K (the "good year" for contrast): FY2025 net income $1,260.3M (diluted EPS $4.45), down 51% from FY2024's $2,579.1M — and that decline happened in a year revenue grew 9%. The culprit again was below-the-line: a $528.9M loss on crypto held for investment in FY2025 vs a $687.1M gain in FY2024 — a ~$1.2B swing — plus $356.1M of "other operating expense" and the $311.2M Data Theft cash cost. The tell of this whole business: net income is dominated by mark-to-market on its own crypto and by volume, not by durable operating earnings.

Lens 6 · Earnings Calls (sentiment trend)

No transcripts on disk (transcripts/ empty); reconstructed from shareholder letters and call coverage ``.

• Tone arc (FY2025 → Q1 2026): management has pivoted from "record year / S&P 500 / Deribit" triumphalism (mid-2025, around the July ATH) to a defensive "we will dynamically adjust our expense base" posture. The FY2025 10-K MD&A explicitly says they "plan to dynamically adjust our expense base... increasing or decreasing it as needed" and guided Q1 2026 tech+G&A "in line with" Q4'25 and S&M "in line with or lower" — i.e. cost discipline language that wasn't there at the top.
• What they keep saying: "Everything Exchange," "#1 financial app in the world," "bring the world onchain," stablecoins/USDC as the long-term engine. Armstrong's 2026 roadmap (May 2026) is an 8-point plan led by the global Everything Exchange (crypto + equities + prediction markets + commodities, across spot/futures/options).
• What they stopped emphasizing: retail spot trading as the growth driver — it has been quietly reframed as a mature, cyclical base while derivatives + stablecoins + onchain carry the forward narrative.
• Sentiment read: management conviction in the strategy is high and consistent; honesty about the cyclicality of the core has improved (they now lead with diversification because spot is structurally soft). The risk is narrative inflation — 21 new products announced (tokenized stocks, US perps, AI advisor, even Fannie Mae mortgages) while the P&L is in a loss-making trough.

Lens 7 · Comps

Peer set = US-listed crypto/fintech-trading platforms (no direct pure-play global peer; Binance is private). Multiples are , June 2026; market caps approximate. COIN's own market cap and TTM EPS are , the rest `` where from filings.

Read: On NTM P/E, COIN (~70x) trades at a clear premium to Robinhood (~46x) — the market is paying up for the regulatory moat + the Everything-Exchange option value, and pricing in a crypto-volume recovery. 5-year-avg ROE and a clean EV/EBITDA are n/a (COIN's GAAP earnings are too crypto-mark-distorted for a meaningful trailing EV/EBITDA, and the empty CSVs preclude a research-layer derivation). The honest comp statement: COIN is the most expensive name in its peer group on forward earnings, at the trough of its own earnings cycle — you are paying a premium multiple on depressed numbers.

Lens 8 · Stock-Price Catalysts (what moves COIN >5%)

``. The 5-year tape says COIN is a crypto-beta + regulatory-headline instrument, not an earnings-quality compounder:

• Apr 2021 — direct listing, opened ~$381, instant volatility.
• 2022 — fell with the crypto winter / FTX collapse to sub-$40; the −90% drawdown.
• June 2023 — SEC sues Coinbase (securities-law complaint) — sharp drop.
• Feb 27, 2025 — SEC drops the case with prejudice — major positive re-rate.
• May 19, 2025 — added to the S&P 500 — index-inclusion pop.
• July 18, 2025 — all-time high $444.65 — the top, coincident with peak crypto sentiment + GENIUS Act signing (July 18, 2025).
• Aug 2025 — Deribit close + convertible-note raises.
• Q4 2025 / Q1 2026 prints — back-to-back losses + volume collapse → grind down to the 52-week low $139.36, now ~$163 — roughly −63% from the ATH in ~11 months.
• Apr 2, 2026 — OCC conditional trust-charter approval — a positive that the falling tape largely shrugged off (crypto sentiment dominated).

Pattern: the market reacts overwhelmingly to (1) crypto spot prices/volumes and (2) binary regulatory events. It reacts least predictably to the diversification story — management's biggest product push "barely moved the stock". COIN is a leveraged way to be long the crypto cycle with a regulatory-optionality kicker; it is not yet priced or traded as a stable cash-flow franchise.

---

Phase C — Judge people & books

Lens 9 · Management

• Brian Armstrong — co-founder & CEO (since 2012). Track record: built the category-defining US crypto exchange from zero to an S&P 500 company; navigated two crypto winters and a three-year SEC war to a dismissal — genuinely impressive franchise-building and regulatory endurance. Skin in the game / control: Armstrong and his trusts hold super-voting Class B stock and collectively control key corporate decisions — this is a founder-controlled company. Founder archetype, not a caretaker — high conviction, high willingness to spend on long-dated bets (Base, Deribit, the 21-product Everything Exchange).
• Emilie Choi — President & COO (joint CODM with Armstrong); Alesia Haas — CFO; Paul Grewal — Chief Legal Officer (the public face of the regulatory fight). A stable, experienced senior team.
• Capital-allocation history — mixed, and the most important judgment call:
• Bold M&A: Deribit ($4.295B total: $721.5M cash + $3,573.1M stock) — strategically coherent (buys the #1 options book) but expensive, largely stock-funded at/near the July-2025 highs, and it added $2.82B of goodwill + $1.39B of intangibles, lifting total goodwill to $4.17B (14% of total assets). Echo ($176M) is a smaller onchain-capital-markets tuck-in.
• Buybacks: repurchased ~3.0M shares for $790.2M in FY2025 and ~6.3M shares for ~$1.1B in Q1 2026, with the board adding a $2.0B buyback authorization in Jan 2026. Buying aggressively into a loss quarter is defensible (stock down 60%+) but optically jarring while running GAAP losses and issuing debt.
• Balance-sheet crypto: holds ~$2.0B of crypto for investment — the source of the earnings volatility (Lens 5). A deliberate "eat our own cooking" stance that makes reported EPS a partial proxy for BTC/ETH price.
• Red flags (governance): the dual-class founder control is the structural one; comp includes a famous CEO performance-option mega-grant (Aug 2020) vesting on stock-price hurdles. Related-party customer revenue is immaterial ($9.6M FY2025) — clean there.

Verdict on people: A-grade founder/operator and franchise-builder; capital allocation is aggressive and somewhat pro-cyclical (stock-funded M&A and buybacks at the top of the cycle). You are backing Armstrong's long-game vision — which is the bull case and, if crypto stays soft, the bear case.

Lens 10 · Forensic Red Flags

Forensic lens. Every figure labeled.

• Earnings quality — the headline issue. GAAP net income is dominated by non-operating mark-to-market on Coinbase's own crypto ($528.9M loss FY2025; $482.4M loss Q1 2026 ). The operating business is more stable than net income suggests, but management's preferred Adjusted EBITDA ($2,808M FY2025) strips out crypto marks, SBC, and the Data Theft costs — flattering. Watch the gap: GAAP net income $1,260M vs Adj. EBITDA $2,808M is a ~$1.5B wedge.
• Stock-based compensation: $839.4M FY2025 (~12% of revenue) — high, though notably down from $912.8M FY2024. But Q1 2026 SBC rose to $248.1M (from $190.7M) even as revenue fell 31% — SBC growing into a downturn is a dilution/quality flag.
• Goodwill & intangibles: $4.17B goodwill + $1.40B intangibles post-Deribit (vs $1.14B + $0.05B a year earlier). No impairment yet, tested annually Oct 1 — but Deribit was bought at the top; if derivatives volumes disappoint, an impairment is the canary.
• Cash vs earnings: operating cash flow ($2,426.4M FY2025; $182.7M Q1 2026) tracks below the "good" GAAP years and was hit by the $311.2M Data Theft cash outflow — a real economic cost, not a paper item.
• Balance-sheet leverage: ~$7.28B total indebtedness at YE2025 (Senior Notes $1.74B + four convertible-note series: 2026 $1.27B, 2029 $1.50B, 2030 $1.27B, 2032 $1.50B), with $1.27B of 2026 converts now in current liabilities. Manageable against $10.2B cash, but the converts dilute on conversion and the leverage grew via the Aug-2025 raises.
• Customer-asset off-balance-sheet: $376.1B of customer crypto held with full keys, recognized as an unrecognized obligation — standard for a custodian, but it means a custody failure is an existential, off-B/S tail risk (the Data Theft Incident is the live reminder).

Data Theft Incident (operational red flag). Disclosed via 8-K May 15, 2025: a threat actor bribed overseas support contractors to obtain customer data, then ran social-engineering attacks. No passwords/private keys compromised; $311.2M cash cost in FY2025 (voluntary reimbursements + legal) with ongoing reputational, legal, and governmental-investigation exposure.

Regulatory findings (required sub-section).

• SEC EDGAR EFTS (LR + AAER): regulatory/regulatory-findings.md reports 0 Litigation Releases and 0 AAERs naming Coinbase Global since 2021.
• The SEC v. Coinbase civil suit (filed June 2023, alleging operation of an unregistered exchange) was dismissed with prejudice on Feb 27, 2025, with the SEC stating the dismissal was "not based on any assessment of the merits" — a policy reversal under the new Crypto Task Force, not a courtroom loss. Net regulatory posture has shifted from existential threat to tailwind (CLARITY Act + GENIUS Act passed 2025; OCC conditional trust charter Apr 2026).
• 10-K Item 3 / Note 21 (company's own disclosure): (1) Underwood v. Coinbase Global — purported securities class action; Securities Act §12(a)(1)/§15 and state claims survived appeal and are proceeding to discovery (Feb 2025); names Brian Armstrong personally; outcome uncertain, no estimable loss. (2) State securities regulators (10 states) re: staking — several (Alabama, Kentucky, Illinois, South Carolina, Vermont) dismissed/withdrawn in Mar–Apr 2025; California, New Jersey, Wisconsin, Maryland settled to halt new staking pending adjudication. (3) Ongoing investigative subpoenas re: asset-listing, staking, and stablecoin/yield products.
• Conclusion: No SEC enforcement findings (EFTS LR/AAER = 0); the marquee SEC action was dropped with prejudice. Residual material litigation is the Underwood class action + tail state-staking matters, both company-disclosed, neither currently estimable. Verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-K Item 3/Note 21 as of 2026-06-21.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection

Build from the latest actuals. FY2025 diluted EPS $4.45 on net income $1,260.3M; Q1 2026 ran $(1.49). The defining variable is crypto trading volume (un-forecastable), so the band is wide. Diluted share count ~265–287M; I use ~270M for forward EPS.

Street anchor: consensus FY2026 revenue ~$7.0B, FY2026 EPS ~$3.01 (cut 41% in 30 days), with FY2027 higher YoY.

• Bear (crypto stays soft, Q1 is the pattern not the floor): Transaction revenue annualizes near the early-Q2 ~$215M/quarter run-rate → FY2026 total revenue ~$5.5–6.0B; crypto marks stay negative; operating leverage works in reverse. FY2026 EPS ~$0.50–1.50. FY2027 EPS ~$2–3 on a mild recovery. FY2028 EPS ~$3–4.
• Base (volumes stabilize, S&S + derivatives grow, no big crypto crash): roughly the Street path. FY2026 EPS ~$3.00, FY2027 ~$5.00, FY2028 ~$6.50.
• Bull (next crypto up-cycle + Everything-Exchange take): spot volumes re-accelerate, derivatives + tokenized assets + stablecoin payments compound, crypto marks swing positive. FY2026 EPS ~$5, FY2027 ~$9, FY2028 ~$13+.

At $163, the base case ($5 FY2027 EPS) is a ~33x forward P/E two years out — still rich for an earnings stream this volatile. The price is discounting the bull cycle, not the base.

No forecast.ts create logged — per --watchlist rules (breadth mode banks dossiers only; a Brier forecast is logged only on genuine commitment, which is a human-gated /thesis step). The base call to track when promoted: "COIN FY2026 non-GAAP EPS ≥ $3.00", resolves 2026-12-31.

Lens 12 · Bull vs Bear

Bull case. Coinbase is the regulated, trusted, S&P-500 front-door to crypto at the exact moment US policy turned from hostile to constructive (SEC case dropped, CLARITY + GENIUS passed, OCC trust charter). The moat — trust, custody scale, licensing, Base, and now the #1 derivatives book via Deribit — is real and widening as rules formalize. The revenue base is diversifying away from cyclical spot into recurring stablecoin + derivatives + onchain (S&S +23% in FY2025; derivatives volume +169% YoY; USDC held at ATH $17.8B). If crypto enters another up-cycle, this is a high-operating-leverage call that re-rates violently — the July-2025 $444 ATH shows the ceiling. The "Everything Exchange" is genuine option value: tokenized equities, prediction markets, and stablecoin payments are large adjacent TAMs where Coinbase has distribution and a regulatory head-start.

Bear case (2–3 things that could permanently impair).

1. The core is a low-moat, hyper-cyclical fee on a commodity (spot trading), and it's structurally eroding — Consumer transaction revenue was down in FY2025's bull year and −48% in Q1 2026. If retail trading is in secular decline (take-rate compression + competition + maturation), the highest-quality earnings line shrinks permanently.
2. The diversification engine's crown jewel is a counterparty's product. ~Half of S&S is the Circle/USDC yield, and the GENIUS Act explicitly threatens the third-party-rewards structure — the deal is up for renewal in 2026 and could be rewritten to Coinbase's detriment. Single-counterparty concentration (Circle = 23% of total revenue in Q1 2026) is rising into a regulatory headwind.
3. Earnings are not bankable. Two consecutive GAAP losses, net income swinging $1.2B+ on crypto marks, SBC ~12% of revenue and rising in a downturn, $4.2B goodwill from a top-of-cycle acquisition. The market pays ~70x NTM earnings for this volatility.

Pre-mortem (18 months out, thesis broke): crypto chopped sideways-to-down through 2026–27; spot volumes never recovered; the Circle deal got re-cut under GENIUS, halving the stablecoin yield; Deribit's volumes underwhelmed and forced a goodwill impairment; the stock de-rated from ~70x NTM to a "no-growth financial" ~15–20x on ~$2 normalized EPS → sub-$60. The diversification narrative turned out to be smaller and lower-margin than the spot fees it was replacing.

Are multiples too high? Yes, on any near-term earnings basis. ~70x NTM / ~97x trailing-forward P/E prices a crypto-volume recovery as the base case. The bull case requires the next cycle, on the Street's timing.

Contrarian view (what the market refuses to see, both directions): Bulls refuse to see that the "diversified, recurring" revenue they're paying a fintech multiple for is more regulatorily and counterparty-fragile (GENIUS/Circle) than the spot fees it replaces. Bears refuse to see that the OCC bank charter + formal US market-structure law could let Coinbase become the regulated rails for tokenized TradFi — a far larger prize than crypto trading — making today's depressed earnings the wrong denominator entirely. The truth is binary and cycle-dependent, which is exactly why ~70x at a trough is the wrong entry.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case.

• What structurally breaks the money machine: spot retail trading — the richest-margin line — is both the most competed and the most cyclical surface in the business, and it was already flat-to-down in a bull year. The bull thesis quietly assumes spot is a stable base; the tape says it's a melting one.
• Revenue concentration: Circle/USDC is 23% of total revenue (Q1 2026) and climbing — a single counterparty whose economics the GENIUS Act can legislate away. That is a 20%+ revenue line one regulatory rewrite from impairment. Add: 84% US-concentrated revenue undercuts the "global" growth story.
• Why the moat is weaker than bulls think: the moat is regulatory/trust, which protects share, not take-rate. Coinbase can be the safest venue and still watch fees compress to Robinhood/Kraken/zero-fee levels. Trust ≠ pricing power.
• Most dangerous underestimated competitor: not another crypto exchange — it's Robinhood (crypto + equities + retail distribution at lower fees, the original "everything app" for retail) and, structurally, stablecoin issuers + neobanks that disintermediate the on-ramp. If tokenized equities become real, every brokerage competes for it.
• Worst capital-allocation calls: $4.3B mostly-stock Deribit deal at the cycle top (now $4.2B of goodwill exposed to impairment), and $1.1B of buybacks in a single loss quarter while carrying $7.3B of debt and issuing converts. That's pro-cyclical capital allocation.
• Assumptions that must hold for ~$163: crypto volumes recover on roughly the Street's timeline; the Circle deal survives GENIUS largely intact; derivatives + stablecoin payments scale into real margin; no major security incident repeats. Break any one and the multiple compresses.
• If growth disappoints 20–30%: on a ~$3 trough-ish EPS, a de-rate to a "cyclical financial" 20–25x P/E → ~$60–75, i.e. −55% to −63% from here. The downside is not theoretical — the stock did exactly this from the July-2025 ATH.
• Single scenario that permanently impairs: a custody/security failure that breaks the trust moat (the Data Theft Incident shows the attack surface is live and the cost real — $311.2M) or a GENIUS-driven re-cut of the Circle deal that structurally halves the stablecoin yield. Plausibility: security tail = low-but-nonzero and existential; Circle re-cut = moderate and already in motion.

Lens 14 · Management Questions (ordered by information value)

1. With the GENIUS Act restricting third-party stablecoin rewards and the Circle agreement up for renewal in 2026, what is the realistic range of outcomes for your USDC economics, and how much of the ~$1.35B FY2025 stablecoin revenue is genuinely defensible?
2. Consumer transaction revenue fell 48% YoY in Q1 2026 and was already down in FY2025's bull year — how much of that is cyclical vs. a secular decline in retail spot take-rate, and what is your normalized blended take-rate assumption?
3. You bought Deribit largely in stock near the July-2025 highs, adding $4.2B of goodwill. What derivatives-volume and revenue trajectory justifies that price, and at what point would you take an impairment?
4. You repurchased ~$1.1B of stock in a quarter you reported a $394M GAAP loss while carrying $7.3B of debt — walk me through the capital-allocation logic versus retaining the cash.
5. How much of the "Everything Exchange" (tokenized equities, prediction markets, US perps, AI advisor, mortgages) is real near-term revenue versus optionality, and what's the 2026–27 revenue contribution you're actually underwriting?
6. Your net income swung ~$1.2B on crypto marks in FY2025 — why hold ~$2B of crypto on the balance sheet at all, given it makes GAAP EPS a proxy for BTC/ETH price rather than operating performance?
7. What is the through-the-cycle operating-margin floor you will defend, and which costs are truly variable when volumes drop 30%+ in a quarter?
8. SBC rose to $248M in Q1 2026 even as revenue fell 31% — what is the multi-year path to bringing SBC below 10% of revenue?
9. The OCC conditional trust charter — what new regulated revenue lines (custody for TradFi, tokenized-asset settlement) does it actually unlock, and on what timeline?
10. International is only 16% of revenue despite the "global" strategy — what is the concrete plan and investment to change that, and where?
11. What is your stance on retail spot fees as competition (Robinhood, zero-fee venues) intensifies — do you defend share by cutting take-rate, and what does that do to the model?
12. On the Underwood securities class action that survived appeal and names the CEO personally — what is the realistic exposure and your reserving posture?
13. How do you think about Base monetization — at what scale of onchain activity does it become a material, standalone revenue driver rather than a strategic loss-leader?
14. Bitcoin/ETH treasury, $7.3B debt, $10.2B cash — what is your target liquidity and leverage through a deep crypto winter, and what's your plan if the 2026 converts can't be refinanced cheaply?
15. If crypto volumes stay at the Q1 2026 trough for two years, what does Coinbase look like — which products survive, which get cut, and what's the sustainable earnings base?

---