Coinbase Global, Inc. (NASDAQ: COIN) — The Moat Widened; the Profit Pool Shrunk

Prepared by: Independent Equity Research Analyst desk: Fundamental Equity Research Report date: 2026-08-12 · Coverage: UPDATE — follow-up to the 2026-06-10 report Price (2026-08-11 close): $148.58 · Market cap: $39.2B · Enterprise value: approximately $36.4B 52-week range: $139.11–$402.16 · Q2-end shares outstanding: 263.8M Sector / sub-sector: Financials — Capital Markets / Cryptoasset Platform · CIK: 0001679788

⚡ Claude’s Take

This block is the author’s own independent opinion and is provided for general information only. It is not investment advice. The analytical body that follows (Sections 1–15) takes no position, expresses no recommendation, and carries no price target; valuation there is framed only as embedded expectations and scenarios.

Verdict: HOLD / AVOID NEW MONEY HERE. Conviction: medium. At $148.58, I would underwrite a normal position only around $100–120, hold an existing normal-sized position through roughly $120–170, and stop adding above $170 unless Coinbase proves that stablecoin, custody, derivatives and new products can grow through a falling crypto tape. The prior June zone in the mid-$120s was too generous: the bull falsification test has now triggered because Subscription & Services declined in a second consecutive weak quarter. The stock is cheaper, but the evidence supporting annuity-like earnings is weaker.

Tag: “The moat widened; the profit pool shrank.”

Coinbase remains the best-positioned regulated crypto platform in the United States. It has a trust and custody advantage, direct regulatory relationships, the leading USDC economics, derivatives liquidity through Deribit, institutional financing, Base, and an increasingly broad “Everything Exchange.” Q2 improved several structural facts: institutional transaction revenue rose 65% with Deribit, management said the Circle agreement renewed on unchanged terms, global trading share reached a company-record 10.3%, and regulators enabled more cross-border derivatives activity. The franchise did not break.

But the income statement did. Q2 revenue fell 19%, consumer transaction revenue fell 31%, Subscription & Services fell 12%, operating loss reached $113M, and net loss was $359M. S&S rose to 48% of net revenue only because transaction revenue fell faster. Stablecoin revenue declined despite larger balances because lower rates overwhelmed growth; blockchain rewards fell 42%; and Coinbase stopped presenting aggregate Trading Volume as a key metric just as spot volume fell 38%. Adjusted EBITDA of $208M required adding back $238M of stock compensation, $52M of restructuring, and volatile crypto marks. That is diversification, but it is not de-cyclicalization.

The correction to June’s buyback analysis matters. Coinbase did not repurchase $1.9B at roughly $303 per share. The prior report divided cumulative dollars by Q1-period shares. The correct inception-to-date figure through Q2 is about 10.1M shares for $1.97B, or roughly $195; H1 purchases averaged approximately $167, and Q2 purchases averaged $148.94. Management did buy heavily into weakness and reduced shares despite stock compensation. That improves stewardship, although the proprietary $1.47B crypto investment book, $5.46B of goodwill/intangibles, founder control, and large time-based equity awards keep the verdict from becoming clean.

This is a falling-knife / high-beta reset, not a proven compounder at a temporary discount. COIN is below its 21-, 50-, and 200-day moving averages; the factor model clusters it with Bitcoin products and MSTR, not with CME or ICE; and the shares remain almost 3× book and about 6.2× trailing sales despite a trailing GAAP loss. The bullish flip is two quarters in which stablecoin, custody, derivatives and new products grow while consumer spot remains soft, with positive operating income after stock compensation. The bearish flip is another quarter of simultaneous transaction and S&S contraction, especially if USDC reward competition rises or the Circle concentration worsens.

📈 Stock Price Action — Five-Year Event Map

COIN has traded less like a toll road and more like a levered option on crypto liquidity. Since its April 2021 direct listing, the shares have moved from a $429.54 intraday high to a $31.55 low, recovered to $444.65, and fallen back to $148.58. That history is the empirical reason valuation cannot be separated from cycle position.

The current price is 63.1% below the 52-week high and 6.8% above the low. It fell 6.6% over 30 days, 26.4% over 90 days, and 53.5% over twelve months, although it remains 5.3% above its February level. The shares sit below the 21-day EMA ($156.12), 50-day EMA ($162.87), and 200-day EMA ($200.29). FactorsToday reports a 2.58 realized beta, approximately 42% annualized specific volatility, and 66.5% model-explained variance. This is a high-beta consolidation close to the low, with no evidence yet of a durable trend reversal.

Changes Since the 2026-06-10 Report

• The June bull test failed. Subscription & Services declined alongside transaction revenue for a second consecutive weak quarter. Q2 S&S fell 12% and transaction revenue fell 22% year over year.
• The buyback verdict improves after correction. The June report paired cumulative repurchase dollars with period-only shares. Corrected cumulative average cost is approximately $195, H1 2026 approximately $167, and Q2 $148.94—not approximately $303.
• Circle renewal risk receded. CFO Alesia Haas said the revenue-sharing agreement satisfied its renewal conditions and continued on the same terms. Counterparty concentration nevertheless rose to 26% of Q2 revenue.
• Derivatives strengthened the platform. Institutional transaction revenue rose 65%, largely because of Deribit, and CFTC staff provided relief enabling parts of Coinbase’s foreign perpetual and collateral-transfer model.
• The legislative clock slipped. Senate Banking advanced the CLARITY Act 15–9, but it had not become law by August 12. The June report’s late-summer enactment expectation was not met.
• Reporting visibility weakened. Coinbase stopped treating Trading Volume as a key metric, arguing that multi-asset economics made aggregate volume less useful. The rationale is plausible; the timing, during a 38% spot-volume decline, is adverse.
• The balance sheet changed. Corporate cash fell to $8.61B after repayment of $1.27B of converts and $1.24B of H1 buybacks. Debt declined to $5.94B, while the crypto investment book fell to $1.47B fair value.
• The tape weakened. COIN moved from $154.10 at the June report to $148.58 and remains below every major moving average, with 12-month performance of approximately −54%.

1. Executive Summary

Coinbase is the leading regulated US cryptoasset platform, combining retail and institutional trading, custody, stablecoin economics, staking, financing, derivatives, developer infrastructure and an increasingly broad set of exchange-like products. Its competitive position improved during the current downturn: the company reports record global trading share, retains material ETF custody relationships, owns the dominant crypto-options venue in Deribit, renewed its Circle economics, and continues to expand Base and institutional services. Regulation, once purely an existential cost, increasingly functions as a barrier to entry.

The earnings model remains much weaker than the franchise. Q2 2026 total revenue was $1.220B, down 18.5%; transaction revenue was $599M, down 21.6%; and S&S was $555M, down 12.2%. Consumer trading remains the swing factor and fell 30.5%. Institutional transaction revenue grew 64.5%, but much of the increase was acquired with Deribit. Stablecoin revenue fell 5.4% because lower interest rates reduced economics by $55.9M in the quarter despite higher balances. Blockchain rewards fell 42.3%. Three consecutive quarterly losses and a trailing GAAP loss of roughly $988M show that the business is still a crypto-cycle engine with a service layer, not an exchange annuity with incidental crypto exposure.

Quality of earnings is poor at the net-income line and mixed at the operating line. Coinbase holds $1.47B of crypto for investment, creating a $209.5M Q2 mark loss and $691.9M H1 loss. Adjusted EBITDA removes those marks, but also removes stock compensation that was $238.3M in Q2—greater than adjusted EBITDA itself. Cash flow is distorted by custodial asset/liability movements. The cleanest gauges are operating income after stock compensation, revenue composition, native-unit assets, customer activity, and per-share dilution. On those measures Q2 was weak: operating loss $113.5M, MTUs 7.6M versus 8.7M, and AOP down 42%, partially offset by higher Bitcoin units and USDC balances on platform.

The balance sheet is liquid but less pristine after acquisition, repurchases and convert repayment. Corporate cash plus marketable securities was $8.79B against $5.94B of debt. Adding $1.47B of crypto and $0.84B of strategic investments produces approximately $5.15B of net identified financial assets, but goodwill and intangibles total $5.46B. Tangible equity is about $7.62B, or $28.90 per share. Liquidity risk is low; impairment, mark volatility, and capital-allocation risk are meaningful.

At $148.58, equity value is approximately $39.2B and conventional enterprise value approximately $36.4B. The stock trades near 3.0× book, 5.8× trailing EV/revenue and 6.2× price/sales while trailing earnings are negative. A conservative sum of parts—net identified financial assets plus a mature multiple on current S&S and a cyclical multiple on transaction revenue—supports materially less than the market value unless new products grow rapidly or normalized crypto activity recovers. The institutional body takes no position; it concludes only that current valuation embeds a successful transition from crypto-beta earnings to diversified financial infrastructure that Q2 did not yet demonstrate.

2. Business Overview

What Coinbase owns

Coinbase is better understood as a stack of regulated and technological franchises than as a single exchange:

1. Consumer. Retail brokerage and custody for spot crypto, Coinbase One subscriptions, staking, wallet, card, stablecoin rewards, lending and access to newer products. Consumer transaction revenue remains the highest-margin and most cyclical line.
2. Institutional. Prime brokerage, agency and principal execution, financing, qualified custody, ETF servicing, derivatives and infrastructure. Deribit adds the leading offshore crypto-options liquidity pool.
3. Stablecoin economics. Coinbase receives a share of income earned on USDC reserve assets and earns revenue on USDC held on-platform. This behaves like a balance-and-rate-sensitive revenue share, not software subscription revenue.
4. Blockchain and developer services. Staking, Base sequencing economics, Coinbase Developer Platform, Wallet, payments and x402. These are strategically important but individually less disclosed.
5. New markets. Prediction markets, token sales, decentralized exchange access, perpetuals, commodities and tokenized assets support the “Everything Exchange” strategy.

The model joins valuable trust infrastructure to highly elastic activity revenue. For institutions, regulation, custody insurance/control systems, balance sheet, execution and integration are bundled. For retail users, the bundle is brand, fiat rails, product convenience and perceived safety. For developers, Coinbase offers tools and distribution. The customer can multi-home easily, but rebuilding the entire regulated stack is expensive.

Revenue mechanics

Calling all S&S “recurring” overstates quality. Stablecoin revenue depends on short rates and USDC balances; staking depends on token price and protocol rewards; interest/finance depends on rates, collateral and demand. Custody and subscription revenue are more contractual, but not separately disclosed at enough scale to carry the thesis. The classification diversifies mechanics, yet most drivers remain linked to the same crypto and liquidity regime.

Unit economics and reporting quality

Coinbase’s retail fees remain high relative to institutional execution and mature exchanges. Management does not disclose customer-acquisition cost, cohort contribution margin or retention economics in a form that allows a clean lifetime-value calculation. It says growth marketing targets roughly a one-year payback, and recent cohorts outperformed, but that is management hypothesis rather than audited evidence. USDC rewards are especially important: they can grow balances but are a variable subsidy, not a fixed moat investment.

AOP fell to $245.9B from $425.0B as crypto prices declined, while Bitcoin units on platform rose to 2.9M from 2.8M and USDC AOP rose 19% to $8.87B. Native-unit retention suggests custody trust survived the drawdown. MTUs nevertheless fell to 7.6M. The decision to discontinue Trading Volume as a key metric may reflect a broader mix; it also removes the denominator required to monitor take-rate compression. Investors should reconstruct spot and derivatives activity from revenue, disclosed percentage changes, and external venue data.

3. Industry Dynamics

Demand: secular adoption inside a violent cycle

Crypto ownership, tokenization, stablecoin payments and institutional access are secular demand drivers. Spot ETFs made Bitcoin exposure available without a crypto exchange account; stablecoins increasingly function as settlement assets; tokenized funds and securities bring blockchain rails into traditional finance; and derivatives deepen institutional risk transfer. Those forces expand the addressable pool of assets and services.

They do not guarantee attractive exchange economics. Spot ETFs can disintermediate retail trading while creating custody revenue at lower take rates. Stablecoin volume can grow while reserve yields fall. Institutionalization replaces high-fee retail flow with lower-fee, price-sensitive professional flow. Regulatory clarity enlarges the market but also invites banks, brokers and mature exchanges. Demand growth and industry profitability can move in opposite directions.

Supply: capital floods visible returns

The crypto capital cycle is unfavorable at the transaction layer. Technology is replicable, liquidity is mobile, customers multi-home, and global competitors such as Binance, Kraken, Robinhood, OKX and decentralized venues compete aggressively on price and product. CME, ICE and Cboe bring regulated distribution and existing clearing networks. When volumes and token prices rise, capital and products enter quickly; take rates then compress.

Barriers are higher in custody, licenses, surveillance, banking relationships and institutional integration. Coinbase’s decade of operating history, public-company controls and regulatory perimeter are scarce. Deribit’s options liquidity adds a true network effect: market makers and users prefer the venue with the deepest order book, which attracts more activity. The combined moat is therefore strongest in trusted infrastructure and derivatives liquidity, moderate in stablecoin distribution, and weak in a standalone retail spot trade.

Regulation: moat builder and moat dilutor

The SEC dismissed its enforcement action in February 2025 without deciding the merits. The GENIUS Act became law in July 2025 and created reserve and disclosure requirements for payment stablecoins. CFTC relief in May 2026 supported Coinbase/Deribit’s foreign perpetual and collateral-transfer model. Senate Banking advanced CLARITY 15–9 in May 2026, but the bill had not become law by this report date.

Clear rules help Coinbase because compliance costs become sunk barriers and institutions prefer a regulated counterparty. They also help well-capitalized entrants. Stablecoin rules are specifically two-sided: reserve clarity increases adoption, but issuer-yield prohibitions and possible restrictions on affiliate/third-party rewards could raise the cost of Coinbase’s USDC incentive model. The regulatory moat is real, but it does not confer a monopoly.

Capital-cycle conclusion

Returns should migrate away from commoditized spot execution toward scarce regulated infrastructure, balance-sheet services, custody, clearing, derivatives liquidity and distribution. Coinbase is repositioning in that direction. The risk is that it spends on every adjacent product faster than the profit pool matures. Q2’s $1.33B operating-cost base, restructuring charge and falling revenue are early evidence of that mismatch. Industry growth will create value only if Coinbase converts product breadth into customer captivity and per-share earnings before competition commoditizes the new lanes.

4. Competitive Position

Moat map

Against the relevant peer set

Versus Robinhood, Coinbase has deeper crypto infrastructure, custody and institutional credibility; Robinhood has better consumer engagement, lower-cost operations and a broader household-finance product loop. Robinhood’s Q2 revenue and operating income grew while Coinbase contracted, but Robinhood remains heavily transaction and rate sensitive. The competition is increasingly about primary-account status rather than a single trade.

Versus Binance/Kraken/global crypto venues, Coinbase has superior US regulatory trust and public disclosure but generally higher retail pricing and lower global spot share. Deribit improves product breadth and institutional liquidity. Offshore rivals retain price, leverage and listing-speed advantages.

Versus CME/ICE/Cboe, Coinbase offers native crypto custody, onchain rails and a larger direct crypto user base. Mature exchanges possess stronger clearing network effects, more diverse asset-class volume and substantially more stable margins. Coinbase should not receive their infrastructure multiple until S&S and derivatives support profits through a crypto downturn.

Versus Galaxy and crypto treasury proxies, Coinbase owns a much better operating franchise. Factors still cluster COIN with Bitcoin ETFs and MSTR, however, showing that the equity market has not separated the infrastructure value from asset-price beta.

Greenwald test: can a focused entrant reproduce the advantage?

A new retail app can reproduce spot trading. It cannot cheaply reproduce Coinbase’s licenses, custody history, USDC relationship, institutional integrations, brand and derivatives liquidity simultaneously. That gives Coinbase a real local monopoly over a bundle, not over each component. The economic moat will be visible when customers accept stable pricing and the fixed compliance/technology base yields positive through-cycle operating margins. Current evidence supports the strategic moat but not yet the earnings moat.

Competitive verdict

Coinbase’s position is stronger than its reported quarter and weaker than the “crypto toll road” label. It is the best regulated bundle in US crypto, with a widening product perimeter and real trust assets. Yet customers multi-home, fee pressure is structural, stablecoin economics are concentrated, and much of institutional growth was purchased. The moat protects relevance and survival; Q2 shows it does not guarantee returns on incremental capital.

5. Growth History and Forward Opportunities

History: two peaks, no stable trend line

Revenue in FY2025 remained below FY2021 despite a much larger asset class, more products, institutionalization, and five years of spending. That is the cleanest warning against extrapolating peak quarters. The mix improved—S&S became a much larger part of net revenue—but consolidated economics still followed the crypto cycle. FY2024’s rebound and FY2025’s growth did not produce a durable upward margin path; H1 2026 moved back into an operating loss.

Growth engines

Derivatives. Deribit is the most credible large growth engine. Options liquidity is a network business, institutional customers value portfolio margin and cross-product collateral, and CFTC relief makes the combined venue more useful. Coinbase disclosed more than $4.2T of trailing derivatives volume and said its derivatives activity held roughly flat sequentially while the broader market fell 12%. The limitation is acquisition economics: the $4.30B purchase created approximately $4.2B of goodwill and intangibles, and Q2 institutional revenue of $100M still represents a modest disclosed quarterly return relative to the purchase price. Revenue contribution is not profit contribution.

Stablecoins and payments. Average USDC held in Coinbase products reached approximately $20B, up 44% year over year, and more than 30% of circulation was held in Coinbase products at quarter-end. Renewal on unchanged Circle terms improves visibility. Coinbase is broadening distribution to PYUSD, USDT, EURC and other coins, potentially becoming a neutral stablecoin shelf. The opposing forces are rates, rewards and partner sharing: lower rates cost $55.9M in Q2 stablecoin revenue, and distribution growth requires passing economics to customers and venues. Volume can grow faster than contribution profit.

Prediction markets. Revenue rose 106% sequentially and exceeded a $100M annualized run rate. Management says prediction customers are incremental and cross-trade spot rather than cannibalizing existing activity. This is strategically useful because event contracts create engagement outside crypto price cycles. The evidence is early, disclosure is management-defined, and Robinhood already generated $156M of Q2 event-contract revenue. Product-market fit exists; durable economics and regulatory stability are unproven.

Borrow/lend and institutional financing. Average borrow/lend balances rose to $1.49B from $199M a year earlier. Financing can deepen captivity because collateral, custody and trading integrate naturally. It also adds credit, liquidity and spread risk. The business should be judged on net interest margin, collateral losses and risk-weighted return—not balance growth alone.

Base, developer infrastructure and agentic payments. Base offers low-cost settlement, a developer ecosystem and sequencing economics; CDP and x402 aim to make stablecoins usable for machine-to-machine payments. A growing ecosystem can create complement-based network effects for Coinbase Wallet, USDC and developer tools. Financial disclosure is insufficient to measure the ecosystem’s standalone revenue, subsidy or return on capital. It remains valuable optionality, not a base-case earnings stream.

Tokenization and the Everything Exchange. Tokenized assets, commodities, perpetuals, token sales and decentralized trading expand the product perimeter. The strategic logic is sound: a customer already onboarded and screened can use one balance and identity across more markets. The capital-cycle risk is equally clear—Robinhood, Stripe, banks, exchanges and offshore venues are building overlapping products. Breadth creates value only if it lowers acquisition cost, improves retention and raises contribution profit per customer.

Growth quality verdict

The company is gaining product breadth and trading share in a shrinking market. That is better than losing share, but it does not by itself prove growth quality. Two consecutive quarters of simultaneous transaction and S&S decline satisfy the prior report’s bull falsification test. The next proof point is not another market-share slide; it is positive operating income after stock compensation in a quarter where consumer spot remains weak.

6. Financial Quality

Current P&L

The P&L shows negative operating leverage. Total operating expenses were $1.334B in Q2, above total revenue. Technology and development was $472.8M, sales and marketing $239.8M, general and administrative $356.9M, and restructuring $52.4M. The May reduction removed about 700 roles and reduced headcount to 4,321 from 4,988, but the quarter still carried the cost of a platform built for higher activity and many simultaneous initiatives.

Adjusted EBITDA is not owner earnings

Q2 adjusted EBITDA of $207.8M begins with a $359.5M GAAP loss and adds back crypto marks, taxes, interest, depreciation/amortization, $238.3M of stock compensation, and $52.4M of restructuring, among other items. Removing a volatile non-operating crypto mark is analytically reasonable. Removing recurring stock compensation is not: the awards transfer value and require repurchases to prevent dilution. In Q2, SBC exceeded adjusted EBITDA. The economically useful hierarchy is:

1. Operating income after SBC for the core platform.
2. Operating cash generation adjusted for custodial flows and working capital.
3. Per-share change after buybacks and equity issuance.
4. Crypto investment marks shown separately as capital-allocation outcomes.

On that basis, Q2 produced an operating loss, not positive owner earnings.

Crypto marks obscure reported earnings

Coinbase’s crypto held for investment generated a $209.5M Q2 loss and $691.9M H1 loss. The book had $1.823B of cost and $1.468B of fair value at June 30, including 17,311 BTC worth $1.013B and 150,279 ETH worth $235.7M. A 50% move in the portfolio would change pre-tax value by approximately $734M. These marks are outside operating income but inside net income, making GAAP EPS a poor measure of the franchise and a valid measure of management’s balance-sheet risk appetite.

Trailing net income is approximately negative $988M using the four filed quarters through Q2. The AZI valuation feed instead reports positive $2.63 trailing EPS, which does not reconcile and is excluded. Any screen showing a current P/E is using stale or mismatched data.

Cash flow and custodial distortion

H1 operating cash flow was $380M, down from $1.09B. Coinbase’s cash-flow statement includes large movements in customer custodial funds and corresponding liabilities; these can inflate or depress reported operating cash without changing corporate economics. Capital expenditure is also absent in some aggregator feeds, causing “free cash flow” to equal operating cash flow mechanically. The filing, not the vendor label, controls.

Corporate liquidity remains substantial. Cash fell by $2.67B from year-end because the company repaid $1.273B of 2026 converts and spent $1.243B of cash on H1 repurchases, alongside operating and investment movements. This is an intentional deployment of liquidity rather than a solvency problem.

Balance sheet

Netting cash and marketable investments against debt gives approximately $2.84B of conventional net cash. Including crypto and strategic investments raises identified net financial assets to approximately $5.15B, before taxes, haircuts or other obligations. Goodwill and intangibles equal 42% of equity; a weak Deribit outcome would create impairment risk without directly consuming cash.

Dilution and return on capital

SBC was $839M in FY2025 and $238M in Q2 2026. The Deribit acquisition also used a large stock component. Repurchases have more than offset recent issuance at the gross share-count level, but that result cost nearly $2B of cumulative cash. Return on capital is not stable enough for a precise point estimate: peak-cycle operating margins exceed 30%, trough margins turn negative, and the acquisition denominator expanded materially. The defensible conclusion is that Coinbase has not yet demonstrated a through-cycle 15%–25% return on invested capital after SBC and acquisitions.

Financial-quality verdict

The balance sheet is strong; current earnings quality is weak. Operating income is cyclically negative, adjusted EBITDA excludes a recurring economic cost, net income is dominated by crypto marks, cash flow contains custodial movements, and the acquisition balance introduces impairment risk. Native-unit retention and liquidity reduce survival risk, but they do not make current profitability durable.

7. Capital Allocation

Deribit: strategically sound, financially unproven

Coinbase acquired Deribit on August 14, 2025 for approximately $4.295B—$721.5M cash and $3.573B of Class A stock—with an indemnity escrow. The asset is strategically coherent: derivatives deepen institutional captivity, options liquidity has network effects, and equity was sensible currency during a strong share-price period. Q2 institutional transaction revenue growth and CFTC relief validate part of the strategic thesis.

The return case remains open. Purchase accounting created about $2.82B of goodwill and $1.35B of customer-relationship intangibles. Institutional transaction revenue was $100.1M in Q2, and Coinbase does not disclose Deribit’s full standalone contribution profit. Paying a large multiple for a cyclical venue can be value-creating only if liquidity share persists and cross-margin/custody synergies arrive without fee compression.

Buybacks: corrected from the June report

The June report’s approximately $303 average-cost claim was wrong. It divided cumulative repurchase dollars by only Q1-period shares. Matching the date bases produces the following:

The correction materially improves the verdict. Management accelerated purchases after the shares fell and used Q2 prices close to the current quote. Repurchases reduced gross shares despite SBC. Approximately $1.97B remained under the $4B authorization at quarter-end. The remaining questions are whether buybacks continue only below assessed value and whether the company can fund them from owner earnings rather than balance-sheet drawdown.

Crypto treasury

The $1.47B crypto investment book amplifies the same risk factor already embedded in revenue. Coinbase added $166.6M and disposed of $33.9M in H1. The portfolio can be defended as ecosystem alignment and strategic inventory, but a public shareholder can buy Bitcoin separately. From a capital-allocation perspective it reduces transparency, consumes risk capacity and makes reported earnings more volatile without deepening customer captivity.

Debt and liquidity

Issuing zero/low-coupon converts during strong markets was opportunistic. Repaying the $1.27B 2026 converts removed a maturity and reduced debt to $5.94B. The company has no near-term liquidity problem. The relevant discipline test is how much of the remaining cash cushion is committed to buybacks, acquisitions and the crypto book before the core business returns to positive post-SBC operating income.

Compensation, insiders and governance

Founder Brian Armstrong controls approximately half of voting power through super-voting shares and Coinbase is a controlled company. Other senior executives receive large time-based equity awards rather than predominantly return-on-capital-based awards. This aligns pay with the stock price but not necessarily with per-share intrinsic value through the cycle.

The five-year Form 4 census found no open-market purchase after May 2023. Fred Ehrsam and Tobias Lütke were the only open-market buyers in the period. Code-S sales were overwhelmingly associated with Rule 10b5-1 plans: Armstrong’s parsed five-year sales totaled about 5.10M shares / $1.31B, including 3.24M / $1.04B in the last 24 months. Scheduled sales are weaker signals than discretionary sales, but the absence of buying during the current drawdown is still informative.

Capital-allocation verdict

Capital allocation is mixed, upgraded from the June report’s mixed-to-poor assessment. The corrected repurchase history and convert issuance show real counter-cyclical discipline; Deribit is strategically credible. The negatives are a large unproven acquisition premium, a proprietary crypto book, recurring SBC, founder control and no recent insider buying. Per-share stewardship improved, but the hurdle remains post-SBC returns on the expanded capital base.

8. Changes and Headwinds — Last Two Years

Franchise positives. The SEC enforcement action was dismissed, the GENIUS Act became law, Coinbase entered the S&P 500, Deribit expanded derivatives capability, USDC balances grew, CFTC relief enabled new products, and management reports substantial trading-share gains. These events lower existential risk and widen the addressable product set.

Earnings negatives. Revenue rolled over, S&S proved correlated with the down tape, rates hurt stablecoin and corporate interest income, blockchain rewards fell, and operating leverage turned negative. Three consecutive GAAP losses and the restructuring show that the cost base and revenue opportunity are not yet synchronized.

Organizational change. Coinbase reduced approximately 700 roles in May 2026, incurred $52.4M of Q2 restructuring cost, and described an AI-native operating model. Chief Legal Officer Paul Grewal and Chief People Officer Lawrence Brock subsequently announced transitions with successor/advisory arrangements. These are not evidence of crisis, but two control-function changes during regulatory expansion and integration raise execution risk.

Disclosure change. Discontinuing Trading Volume as a key metric weakens the ability to separate volume, mix, acquired contribution and take-rate compression. A more diversified platform needs more product-level disclosure, not less.

Regulatory timing. Agency policy is enabling real products, but statute lagged. CLARITY passed committee, faced opposition to later text, and its initial floor timing slipped into September. That lowers near-term certainty and leaves policy more reversible than the June report assumed.

Capital structure. Cash declined, the nearest convertible was repaid, debt fell, buybacks continued at lower prices, and goodwill/intangibles remained elevated. The balance sheet is still a source of strength; its margin of safety is being spent.

The two-year conclusion is asymmetric: Coinbase’s probability of remaining a central crypto institution increased, while confidence in smooth earnings decreased. The franchise got safer; the stock’s required earnings transition got harder.

9. Risk Analysis

Catastrophic risk. A custody breach, customer-asset shortfall or durable loss of regulatory permission could permanently destroy the trust franchise. A prolonged crypto winter is more probable but less terminal: it would likely produce further operating losses and multiple compression while the balance sheet funds the transition.

Risk interaction. The most damaging plausible combination is not one isolated event. It is lower crypto activity plus lower rates, causing transaction and stablecoin revenue to fall while rewards, SBC and integration costs remain high; a regulatory delay then prevents new derivatives/products from filling the gap. The current balance sheet can survive that scenario, but current valuation may not.

10. Valuation Discussion — Embedded Expectations

This section expresses no recommendation and no price target. It frames what the current market value requires and how alternative operating assumptions change the implied value distribution.

Current metrics

At $148.58 and 263.8M Q2 shares, market capitalization is approximately $39.2B. Adding $5.94B of debt and subtracting $8.61B of cash and $0.17B of marketable investments gives conventional enterprise value of approximately $36.4B. Trailing revenue through Q2 is approximately $6.28B, so EV/revenue is 5.8× and price/revenue 6.2×. Price/book is 3.0× and price/tangible book approximately 5.1×. Trailing GAAP earnings are negative, so P/E is not meaningful.

The own-history feed shows lower P/B and P/S percentiles, but the comparison is weak: Coinbase has only five years of public history, the business perimeter changed through Deribit, and the feed’s trailing sales/earnings do not fully reconcile to filed quarters. Current filing-derived multiples control.

Component economics

Q2 annualizes to approximately $2.22B of S&S revenue and $2.40B of transaction revenue. Identified net financial assets are approximately $5.15B if crypto and strategic investments are included at reported fair value. A simple reverse sum of parts illustrates the embedded premium:

The market therefore assigns roughly $14B–$19B to future recovery and optionality even after granting healthy multiples to the existing components. That residual can be justified by a return toward FY2024/2025 activity, successful Deribit integration, stablecoin balance growth, Base monetization, prediction markets and tokenization. It is not supported by Q2’s current earnings.

Through-cycle earnings bridge

A revenue multiple is useful only because current earnings are below a normal cycle; it should not become the endpoint. The operating model has four distinct margin structures. Consumer transactions have very high incremental margins but the least stable volume and the greatest fee pressure. Institutional and derivatives activity has lower take rates, but liquidity and custody can make the revenue more durable. Stablecoin revenue carries high gross economics before rewards and partner sharing, then moves with balances and short rates. The remaining S&S and developer lines range from attractive subscription/custody revenue to token-price-sensitive staking. Applying one consolidated margin to all four hides the source of value.

The FY2021 and FY2024 margins establish upside operating leverage, while FY2022, FY2023 and H1 2026 establish the downside. The five-year median is not a clean “mid-cycle” answer because product scope, accounting and acquisition perimeter changed. A better bridge begins with the Q2 cost base: $1.33B of quarterly operating expense against $1.22B of revenue. Restructuring can remove approximately $52M of one-time cost and lower headcount, but ordinary SBC remains. If quarterly revenue returned to $1.75B with expenses held near $1.40B, operating margin would approach 20%; if revenue stayed near $1.25B, another round of cost reduction would be necessary merely to reach breakeven. That illustrates why volume and cost discipline, not the label attached to revenue, drive the equity outcome.

Per-share math adds another hurdle. A business can report higher adjusted EBITDA while issuing equity or spending cash to neutralize awards. H1 adjusted EBITDA less SBC was only $24.7M, while H1 repurchase cash was approximately $1.18B. The repurchases were executed at much better prices than previously calculated, but they were funded primarily from accumulated liquidity rather than current owner earnings. A credible normalized case therefore needs all three: recurring post-SBC operating profit, restrained acquisition/treasury deployment, and a stable diluted share count. Without that bridge, a recovery in headline revenue can still fail to compound value per share.

Finally, the current residual value should be compared with evidence milestones rather than a single forecast. Circle contribution after rewards, Deribit contribution after integration cost, product-level take rates, Coinbase One cohort economics and Base/developer profit are the lines that can convert optionality into a capitalizable stream. Until those are disclosed, the market is underwriting a portfolio of correlated ventures with a strong distribution platform—not a measured collection of annuities.

What the price implies

A normalized post-SBC operating margin of 15% on $8B revenue would produce about $1.2B of operating income. After tax and finance items, current equity value would still represent roughly 35–40× normalized earnings before giving separate credit to financial assets. A 25% margin on $10B revenue produces $2.5B of operating income and makes the valuation more conventional. The difference is not a small modeling choice: the current price requires either a material revenue recovery, much better cost absorption, or both.

Scenario framework

Valuation verdict

The stock is cheaper than at its 2025 peak and expensive relative to current owner earnings. Balance-sheet value limits insolvency risk but does not anchor the equity near $149. The binding question is whether $14B–$19B of implicit growth/quality value becomes measurable post-SBC profit before fee competition and lower rates erode the legacy pools. Q2 moved that proof backward, not forward.

11. Variant Perception

Consensus narrative 1: S&S has de-cyclicalized the model

The strongest bullish framing is that Coinbase now earns nearly half of net revenue outside transactions, so the next crypto winter will resemble an exchange slowdown rather than an earnings collapse. The narrow fact is true: S&S was 48% of Q2 net revenue. The broader inference is not. S&S declined from $694.5M in Q4 2025 to $583.5M in Q1 2026 and $555.1M in Q2. Stablecoin, staking and interest are connected to rates, token prices, balances and activity. Mix rose because transactions fell faster.

Variant: Coinbase diversified away from Bitcoin spot fees, not away from the crypto/liquidity regime. S&S is a shock absorber, not yet an independent flywheel.

Consensus narrative 2: record market share proves a widening moat

Management reports total crypto trading share rising from 1.5% in Q2 2023 to 10.3% in Q2 2026. This is strategically favorable. But the market definition changed, Deribit enlarged the product perimeter, stablecoin conversions can be included differently across competitors, and share fluctuated materially within the period. Under a strict Greenwald test, rapidly changing market boundaries do not prove stable share.

Variant: the cleaner moat evidence is stable custody share and rising native BTC/ETH units, not the headline trading-share series. Trust/custody is stronger than retail execution.

Consensus narrative 3: regulatory clarity creates an exchange toll road

Dismissal of the SEC case, stablecoin legislation and CFTC product relief reduce existential risk. Yet clarity also lets banks, brokers and mature exchanges compete. The profit pool can expand while returns compress. CLARITY also missed the prior report’s expected timetable, leaving much of the framework dependent on agency interpretation and no-action relief.

Variant: regulation raises Coinbase’s survival value faster than it raises pricing power. It is a barrier against small entrants and an invitation to large ones.

Consensus narrative 4: adjusted EBITDA proves trough profitability

Positive adjusted EBITDA in a down tape supports the idea that the 2022 cost reset permanently improved resilience. Q2 adjusted EBITDA was $208M. But Q2 SBC was $238M, and H1 adjusted EBITDA less SBC was only $24.7M. Q2 operating loss after SBC was $113M.

Variant: the platform is near contribution breakeven in the trough, not durably profitable. Repurchasing stock makes SBC economically visible through cash deployment.

Consensus narrative 5: the stock is cheap because it fell 65%

COIN is far below its high, at a lower own-history P/B and near a 52-week low. Those statistics compare a changing business across extreme crypto regimes. At current filings it still trades around 5.8× EV/revenue and 3.0× book with negative trailing earnings, while $14B–$19B of equity value remains attributable to recovery/optionality under reasonable component references.

Variant: the drawdown removed peak exuberance, not the need for a successful earnings transition.

What the market may underappreciate positively

The bear case can also become too simple. Native asset units grew through the drawdown; Deribit has a genuine liquidity network; the Circle renewal removed a major contract cliff; management bought substantially more stock at $149–169 than the June report recognized; and balance-sheet liquidity creates time. A regulated winner can gain durable share during an industry shakeout even while current profits disappoint. The research edge is holding both claims at once: the strategic position improved and the current earnings proof deteriorated.

12. Fact vs. Interpretation Table

13. Open Questions

1. What are product-level volumes and net take rates after Coinbase stopped treating aggregate Trading Volume as a key metric?
2. How much of the 10.3% trading-share gain is organic spot share, acquired Deribit share, perpetuals, or definitional change?
3. What is Deribit’s standalone revenue, contribution profit, capital requirement and customer retention after integration?
4. What is reward-adjusted USDC contribution profit, rather than gross stablecoin revenue, at 100–200bp lower short rates?
5. What termination rights, pricing resets and partner-sharing obligations exist in the renewed Circle agreement?
6. How many Coinbase One subscribers exist, what do they pay net of waived fees/rewards, and how do retention and contribution compare with nonmembers?
7. Can prediction markets sustain engagement outside election/sports peaks, and what portion cross-sells rather than cannibalizes spot activity?
8. Will Base/CDP/x402 revenue become material enough to disclose, and what incentives are required to support the ecosystem?
9. What portion of the cost reset is durable after restructuring, and which product investments are being stopped rather than merely staffed more efficiently?
10. Will the remaining $1.97B repurchase authorization be deployed only when post-SBC owner earnings and valuation support it?
11. Why maintain a $1.47B directional crypto investment portfolio when operating revenue already supplies large crypto beta?
12. Does CLARITY reach a floor vote, and do amendments preserve third-party stablecoin rewards and the CFTC product pathway?
13. Do control-function departures affect regulatory execution, institutional memory or the product calendar?
14. Can native-unit custody share remain stable if ETFs and competing qualified custodians capture flows?
15. What is normalized through-cycle operating margin after SBC and USDC rewards: negative, mid-teens, or above 20%?

14. What Must Be True

Bull case

The bullish institutional case requires all of the following:

• Regulated custody and Deribit liquidity produce durable customer captivity, not merely acquired volume.
• USDC and partner stablecoin balances grow fast enough to overcome falling rates and reward-sharing leakage.
• S&S returns to growth while consumer spot remains soft, showing that the revenue pool is more than correlated diversification.
• Prediction markets, financing, Base and new assets lift contribution profit per customer without recreating the FY2022 cost base.
• Post-SBC operating margin returns to at least the mid-teens through a normal, not peak, crypto tape.
• Buybacks offset dilution below intrinsic value while the balance sheet retains ample stress liquidity.
• Regulatory pathways survive changes in administration and invite more customer demand than competitor supply.

Bull falsification test: two additional quarters in which transaction and S&S revenue both decline, or S&S remains below $550M while post-SBC operating losses persist, despite product/share gains. That outcome would show that breadth is redistributing a shrinking pool rather than creating independent earnings.

Bear case

The bearish institutional case requires several of these conditions:

• Retail take rates compress faster than volume/share can grow.
• Lower rates and higher rewards shrink USDC contribution, while concentration remains above 20% of revenue.
• Deribit volume grows but acquisition returns disappoint because institutional take rates and partner economics compress.
• Regulatory clarity invites banks, brokers and offshore venues faster than it locks in Coinbase’s incumbency.
• The cost base, SBC and product sprawl prevent positive owner earnings in ordinary activity regimes.
• Crypto-treasury marks and goodwill impairments keep reported returns volatile and below the cost of capital.
• The equity remains priced for an infrastructure transition while factors and earnings continue to behave like crypto beta.

Bear falsification test: two consecutive quarters with weak consumer spot but rising S&S, positive post-SBC operating income, stable or rising native-unit custody share, and demonstrable Deribit/stablecoin contribution growth. That combination would prove the platform can compound without a broad crypto-volume upswing.

Monitoring dashboard

15. Research Scope and Source Summary

This update used the July 30, 2026 Form 10-Q and SEC-filed earnings deck as controlling sources; the FY2025 10-K, Q1 2026 10-Q and 2026 proxy for history, acquisition, governance and compensation; a 60-month SEC census covering all identified Forms 4; company earnings Q&A with numerical reconciliation to filings; SEC, CFTC, White House and Senate materials for regulation; AZI adjusted price history; FactorsToday factor, peer and specific-volatility outputs; and public filings and disclosures from relevant peers including HOOD, CME, ICE and GLXY.

No sector primer was relied upon. The prior published report was used only to identify questions and falsification tests; current claims were revalidated. A third-party transcript copy was incomplete and machine-transcribed, so filings controlled all numerical claims. A market-data provider’s trailing EPS did not reconcile to filed losses and was excluded. Current short interest was not available from a clean, date-stamped public source and is omitted.

Facts, management claims and interpretations are separated throughout. The full dated source list and evidence-quality notes appear in Appendix B.

APPENDIX A — Standard Diligence Questionnaire

Coinbase Global, Inc. (NASDAQ: COIN) — as of 2026-08-12

Supplemental to the research memo. “Fact” refers to filed or dated evidence; “Interpretation” identifies analytical judgment. This appendix carries no recommendation or price target.

General and earnings nature

What is the central investor debate? Whether Coinbase is becoming a regulated financial-infrastructure annuity or remains crypto beta with diversified labels. Q2 supports both sides: trust, share, products and regulatory position improved, while both revenue pools contracted and owner earnings were negative.

Are earnings at a cyclical high or low? (Interpretation) Current earnings are closer to a trough than a high, but the cycle is not demonstrably turning. H1 operating loss was $135M and GAAP loss $754M; price and MTUs are depressed. The business can still become cheaper if low activity persists.

How stable are revenues? Low. Revenue was $7.84B in 2021, $3.19B in 2022, $3.11B in 2023, $6.56B in 2024, $7.18B in 2025 and $2.63B in H1 2026. S&S reduces amplitude but is itself rate/crypto sensitive.

What drives the cycle? Crypto prices, volatility, trading activity, stablecoin balances, short rates, staking economics and regulation. Internal product breadth and costs modulate those external variables; the proprietary crypto book amplifies them.

Business quality and moat

Where is the moat strongest? Regulated custody/trust, institutional integrations, USDC distribution and Deribit options liquidity. Stable custody share and rising native BTC/ETH units are the best quantitative evidence.

Where is the moat weakest? Retail spot execution. Customers multi-home cheaply, fees are transparent, and global/fintech rivals compete on price and product speed.

Do network effects exist? Deribit liquidity and, potentially, Base’s developer ecosystem show network properties. USDC has distribution effects but requires rewards and partner sharing. Coinbase’s consumer app itself is a bundle, not a classic network monopoly.

Is the industry becoming more competitive? Yes. Robinhood, offshore venues, stablecoin issuers, Stripe, banks and mature exchanges are entering overlapping products. Regulation removes small competitors while inviting large ones.

Does growth create value? Only when it improves contribution profit and captivity. Acquired volume, subsidized balances and new-product gross revenue can destroy value if fees/rewards/costs absorb the gain.

Financial condition and accounting

How profitable is the business? Spectacular at cycle peaks and loss-making in troughs. Q2 operating margin was −9.3%. A through-cycle ROIC above the cost of capital has not been demonstrated after SBC and acquisitions.

What is the cleanest earnings measure? Operating income after SBC, adjusted for clearly nonrecurring restructuring but not ordinary compensation; then per-share cash deployment. Net income includes crypto marks, and adjusted EBITDA excludes SBC.

Is cash flow reliable? It requires adjustment. Custodial-fund asset/liability movements distort operating cash flow, and aggregator “FCF” can equal OCF when capex is missing. Filed components control.

How strong is the balance sheet? Strong. $8.61B cash and $0.17B marketable investments exceed $5.94B debt. Broader net identified financial assets are about $5.15B including crypto/strategic investments, subject to haircuts and tax.

What accounting risks matter? Crypto fair-value marks obscure net-income trends; $5.46B goodwill/intangibles creates impairment risk; stablecoin and acquired-product contribution are not sufficiently disaggregated; and custodial presentation complicates cash-flow analysis.

How capital intensive is the company? Physically light, economically heavy. Technology/compliance, SBC, USDC rewards, M&A consideration, liquidity requirements and the crypto book consume capital.

Capital allocation and management

Was Deribit a good acquisition? (Interpretation) Strategically credible and sensibly financed with equity, but financially unproven. The price was $4.295B and standalone profit/return disclosure is insufficient.

Are buybacks creating value? The corrected history is much better than previously reported: cumulative cost approximately $195, H1 approximately $167 and Q2 $148.94. They have offset dilution. Value creation still depends on intrinsic value and sustainable funding.

What is the crypto-treasury policy? Coinbase actively holds BTC/ETH and other crypto for investment; fair value was $1.47B. It aligns with the ecosystem but duplicates shareholder-accessible exposure and magnifies volatility.

How are executives paid? Armstrong takes salary/security costs and no new equity due to large founder ownership; other executives receive substantial time-based equity. Metrics are less tightly linked to through-cycle ROIC than ideal.

What governance risks exist? Founder super-voting control limits minority influence over board, compensation, acquisitions, treasury and repurchases. Two senior control-function transitions add near-term execution risk.

What does insider activity show? No open-market purchase since May 2023. Five-year sales are large but overwhelmingly 10b5-1 plan-associated, making absence of buying the cleaner signal.

Market, valuation and downside

How should the company be valued? By separating S&S, transactions and net financial assets, then explicitly pricing recovery/new-product optionality. P/E is unusable with negative trailing earnings; mature-exchange multiples are premature.

What does the current price embed? Approximately $34B for operating franchises after identified net financial assets. Under reference component multiples, roughly $14B–$19B remains attributable to recovery, new products, network value or higher durable margins.

What can cause permanent impairment? A custody/security loss, adverse durable regulation, Deribit impairment, long-run fee/reward compression, or cost/SBC that prevents returns above the cost of capital.

What is the chance of total loss? Low given liquidity, brand and market position. A realistic severe outcome is multi-year loss-making activity and large equity-value compression, not immediate insolvency.

What recent evidence changed the thesis most? Negative: the second consecutive simultaneous decline in S&S and transactions. Positive: Circle renewal on the same terms, derivatives/regulatory progress, native-unit custody resilience, and the buyback correction.

What single disclosure would be most valuable? Product-level volume, net revenue/take rate, contribution profit and active-customer retention—especially for Deribit, Coinbase One, USDC after rewards and prediction markets.

APPENDIX B — Source Appendix

Coinbase Global, Inc. (NASDAQ: COIN) — Research as of 2026-08-12

This appendix covers the sources used for the August 2026 update, with emphasis on Q2 2026, price action, factor positioning, and the five-year event map. Sources are ordered primary-first. All links were accessed on 2026-08-12 unless stated otherwise.

1. Current Company and SEC Sources — Primary

2. Market, Price, and Factor Sources — Third-Party Quantitative

3. Regulatory and Index Event Sources — Primary

4. Current and Historical Independent Reporting — Secondary

5. Evidence and Contradiction Audit

6. Date-Integrity and Unsupported-Claim Check

• Sources older than 18 months are used only to explain historical price events: the 2022 crypto winter, the 2023 SEC action, and the January 2024 spot-Bitcoin ETP approval.
• The current financial read relies on the July 30, 2026 10-Q/deck and Q&A, not the June 2026 prior report.
• Current price/factor data run through August 11-12, 2026. The All-Factors loading basis is dated July 31; later August 11 nested-model loadings are disclosed only as a freshness check.
• No current short-interest figure was accepted: the prior June report’s 11.7% of float is stale, and available public sources did not provide a clean, date-stamped replacement. The analysis therefore omits short-positioning claims pending a current Nasdaq/FINRA source.
• Claims about prediction-market cross-sell, Coinbase One unit economics, ETF-driven asset outflows, and proprietary market share remain management claims unless independently validated elsewhere in the report.