Report contents
Industry landscape
Where the company competes: US-regulated access in a global price war
Market definition: a licensed brokerage and custodian, not a network. Coinbase sells access. Its core market is the customer-facing layer of crypto spot trading and custody for US retail investors and institutions: the venue that quotes a price, executes the order, holds the asset and produces the tax report. Upstream sit market makers who supply quotes, the blockchains that settle transfers and the banks that move dollars; downstream sit retail traders, asset managers and, increasingly, the sponsors of spot crypto ETFs that use the company as custodian. The margin in that chain accrues to whoever owns the customer relationship and the custody record, which is where Coinbase sits. It earns a percentage of each trade, custody and staking fees, and a distribution payment on reserve income from USDC, the stablecoin issued by Circle.
That structure explains what drives the P&L. Revenue is roughly trading volume times a take rate (the percentage the venue keeps per trade), plus recurring service fees. Volume follows asset prices and volatility, which no exchange controls, so the competitive question is narrower than the revenue line: can Coinbase hold its take rate and its customers while the cost of compliance rises? The 10-K for fiscal 2025, filed with the SEC, lays out the same split between transaction-based and subscription-and-services revenue.
Geography: access is national, price is global. US spot trading has run through a patchwork of state money transmitter licenses and New York's BitLicense, a state-issued crypto permit with its own capital and compliance standards. Europe now sits under MiCA, the EU's Markets in Crypto-Assets regime. Offshore venues such as Binance, OKX and Bybit serve customers outside those perimeters, and Upbit dominates its home Korean market. The practical result is a market that is fenced by licensing but priced by comparison: a US retail customer can see an offshore quote and fee before clicking buy. Access is local; the reference price is not.
Market structure: many venues, low concentration, cheap switching. Crypto spot trading is fragmented across dozens of exchanges plus over-the-counter desks, and no venue's share is large enough to set the global price of bitcoin. Venues do not publish comparable share data, and third-party estimates use inconsistent bases, so a ranked top-five table would imply more precision than exists. It is enough to say that the leaders number well beyond five and that several are offshore and lightly regulated.
What that does to bargaining power is the part that matters. Market makers and large institutions can route orders to whichever venue offers the tightest spread, so institutional take rates run thin everywhere. Retail customers face more friction, but not much: moving assets on-chain is straightforward, and the real switching costs are fiat on-ramps, tax reporting history and plain familiarity.
Entry barriers differ sharply by segment. Offshore, listing a trading pair and renting liquidity is cheap, and competition runs on fees and leverage. Onshore, the barriers are real but administrative: licensing across states, custody controls, banking relationships, surveillance and the compliance payroll that goes with them. A startup cannot amortize that cost across a small book. That is why US-regulated access has consolidated around a handful of names, while the global venue count keeps rising.
What this means for the moat question. Coinbase competes in a submarket where the regulated perimeter is genuinely hard to cross, but where the product inside the perimeter is a commodity priced against offshore alternatives. Scale helps on the cost side, since compliance and technology spending is largely fixed; it helps less on pricing, because the customer can comparison-shop. The open question is whether the licensing and custody position, including the ETF custody relationships, is durable enough to defend the fee line, or whether the fee line keeps compressing as cheaper venues and asset managers absorb more of the customer's economics. That is the test the next section takes up.
Competitive advantage
Does the company have a moat? Moderate, with the durable half in stablecoin distribution
Coinbase's competitive position rates Moderate. Where the protection actually sits is the plumbing around the trading screen: licensing, custody records and, above all, the company's role as the main distribution channel for USDC, the dollar stablecoin that Circle issues. USDC economics produced $292 million of revenue in 2Q26, roughly half of Circle's reserve income, DBS analysts Manyi Lu and Ken Shih estimated in an August 14, 2026 note. That is a fee collected on another firm's float, and it runs through a contract.
A moat means above-normal returns defended by something an entrant cannot simply buy. The defense here is administrative and genuine: US crypto access runs through state money transmitter licenses, New York's BitLicense and, since 2024, the custody arrangements behind spot ETFs. The returns are the weaker leg. Coinbase earned an estimated after-tax return on invested capital (ROIC: operating profit after tax divided by the equity, debt and cash an entrant would have to commit) of about 32% in FY2024 and 11% in FY2025, and its operating profit in the first nine months of 2026 rounded to roughly zero.
US-regulated crypto access is concentrated in a handful of licensed venues and the custody arms of a few institutions. Concentration only matters if the survivors earn excess returns through a full cycle. Coinbase's five-year record shows they can in upswings, and that upswings end.
Market position: 10.3% of global volume and gaining
Coinbase reached a record 10.3% of global crypto trading volume in 2Q26, its third consecutive quarter of gains across spot and derivatives, according to DBS. It is the largest US-listed exchange and the default venue for US institutions that need a regulated counterparty. Ten percent is not dominance: close to nine of every ten trades cleared elsewhere, much of it offshore. Market position: Moderate.
Five-year stability cannot be confirmed; the documented series covers three quarters, all of them gains. The recent direction is favorable, which keeps this above Weak, but three quarters is a trend, not a record.
Returns: 32%, then 11%, then roughly nothing
| USD bn unless stated | FY2021 | FY2022 | FY2023 | FY2024 | FY2025 | 9M 2026 |
|---|---|---|---|---|---|---|
| Revenue | 7.84 | 3.19 | 3.11 | 6.56 | 7.18 | 3.85 |
| Operating income | 3.08 | (2.67) | (0.11) | 2.31 | 1.46 | 0.00 |
| Net income | 3.62 | (2.62) | 0.10 | 2.58 | 1.26 | (0.03) |
| Gross margin (%) | 83.8 | 80.3 | 86.5 | 74.7 | 74.6 | 78.7 |
| Net margin (%) | 46.2 | (82.2) | 3.1 | 39.3 | 17.6 | (0.8) |
| ROE (%) | 98.7 | (44.4) | 1.6 | 31.2 | 10.1 | n.m. |
| Operating cash flow | 4.04 | (1.58) | 0.92 | 2.56 | 2.43 | 0.78 |
ROE is shown for full years only; the 2026 figures are single-quarter returns and do not add. Source: Coinbase reported results.
The ROIC build is short enough to check:
| USD bn | FY2024 | FY2025 |
|---|---|---|
| Operating income | 2.31 | 1.46 |
| Effective tax rate | 12.2% | 17.1% |
| NOPAT (estimated) | 2.03 | 1.21 |
| Equity + total debt − cash | 6.28 | 11.16 |
| After-tax ROIC (estimated) | 32% | 11% |
The swing from 32% to 11% came from two directions at once: operating income fell by more than a third, and the equity base grew by 44% as cash piled up. Neither move says anything about pricing power. Taken together they show a business whose return on capital is set by the crypto cycle and by how much idle capital it carries, not by a stable franchise. FY2021's 98.7% ROE is not comparable: equity was $6.4 billion against $21.3 billion of assets, so a large absolute profit produced an outsized ratio. Core return metrics: Moderate.
Margins: 74% gross and a net margin that misleads
Gross margin has stayed high in absolute terms, at 74.6% in FY2025, but it has narrowed from 86.5% in FY2023 even as revenue more than doubled. The net line is noisier. The third quarter of 2026 showed $0.72 billion of net income against $0.06 billion of operating income, so almost all of the profit came from items below the operating line, most likely gains on crypto holdings rather than fees. The closest large rivals are private or offshore, so no clean margin comparison exists; what is clear is the direction of Coinbase's own. Profitability and margins: Moderate.
Overall rating: Moderate, with protection that outruns the profit
The decisive pair is the record 10.3% share and the collapse in ROIC from 32% to 11% in a single year. The share confirms that barriers exist. The return path confirms they do not yet convert into a dependable premium over the cost of capital, and the first nine months of 2026, with operating profit at zero on $3.85 billion of revenue, underline how thin the through-cycle margin can get. Overall rating: Moderate, matching both sub-ratings; profitability is not strong enough to support a higher label.
Where the advantage comes from and how long it can last: a license wall and a stablecoin revenue split
Three mechanisms support Coinbase's position, and they break in different ways. The licensing perimeter and custody record raise the cost of entering US-regulated crypto. The USDC distribution relationship supplies most of the recurring profit. Scale spreads fixed compliance and engineering spending across a larger revenue base. The metric that tests the whole package is subscription and services revenue as a share of net revenue. DBS put it close to half in 2Q26, against a trading line that offshore rivals price and that can halve in a quarter. If that share keeps climbing while trading stays weak, the moat is working. If it stalls and falls with volume, Coinbase remains an exchange with extra fees attached.
Government intervention: a license wall that stops at the border
Government intervention: Moderate. A US retail account arrives with state licenses, segregated custody, and a tax basis history the customer would have to rebuild elsewhere. For an institution the cost is larger: re-papering custody agreements, retesting controls, and explaining the switch to a board and an auditor. The 2024 spot ETF approvals routed a slice of new US demand through a short list of qualified custodians, and Coinbase is on it. That barrier stops startups; it does not stop JPMorgan.
It is also a barrier drawn by regulators who can move it. The CLARITY Act, working toward a Senate floor vote, would replace a state patchwork with a federal framework. Depending on how it is written, that could cut compliance costs for banks and large brokers and stiffen competition, or entrench whoever holds the first federal licenses; DBS expects the SEC and CFTC to act on their own if the bill stalls. What would break the advantage is a rule separating exchange from custody that forces customers to pick one, or the loss of ETF custody mandates to a bank that bundles custody with prime brokerage. The outcome to follow is whether recurring custody and other service fees keep growing while trading volume shrinks.
Demand-side advantage: stablecoin distribution carries the recurring profit
Customer captivity: Moderate. Retail users are not captive: moving crypto is cheap and comparing a fee takes a minute. Record Coinbase One paid subscribers during the 2Q26 downturn show habit, not lock-in. Institutions are stickier, but they negotiate. The recurring line is not immune either; DBS put its sequential decline at 5%, partly on the timing of USDC payments.
USDC is different in kind. On-platform balances reached a record $20 billion in 2Q26, against $292 million of quarterly USDC revenue. Divide one by the other and Coinbase keeps about 1.5% a quarter, near 6% annualized, on the stablecoin it distributes (estimated). Short-term dollar rates have not been high enough for a pure reserve pass-through to yield that much, so the figure probably blends the reserve split with other USDC-related fees. Whatever the mix, the ratio is the cleanest available gauge of pricing power inside the relationship, because it measures what Coinbase keeps per dollar of stablecoin placed.
The Circle partnership renewed automatically in August 2026 on unchanged terms, DBS reported. That single fact is both the protection and the exposure. Coinbase holds the distribution; Circle holds the issuer economics; the split is renegotiable. If Circle routes distribution through large wallets and exchanges directly, or reopens the revenue share, the recurring line falls even though balances and users do not move. The adverse outcome to track is revenue per dollar of on-platform USDC declining while the balance holds.
Economies of scale: fixed costs that flex with volume
Scale: Moderate. Compliance, licensing, security and engineering are largely fixed, and Coinbase spreads them over roughly $7 billion of annual revenue. The cost base, though, has flexed more than a fixed-cost story predicts. Revenue was almost identical in FY2022 and FY2023, yet the operating loss shrank from $2.67 billion to $0.11 billion after spending was cut. A business that can swing more than $2 billion of cost against flat revenue has a variable cost structure; that protects the downside and gives a smaller entrant less of a penalty for staying small.
Gross margin tells a consistent story. It fell from 86.5% in FY2023 to 74.6% in FY2025 while revenue more than doubled, which points to mix: staking rewards paid to customers run through cost of revenue, and staking grew faster than trading. The filings do not split the contributors, so treat that as an inference. What would break the scale case is compliance spending outrunning revenue as MiCA in Europe and any US federal regime add requirements that fall equally on every licensed venue.
Durability: substitutes tend to arrive as partners
Spot crypto ETFs are the clearest substitute, letting an investor hold bitcoin in a brokerage account without a Coinbase login. Coinbase custodies several of those funds, so it keeps a custody fee while losing the trade, and custody fees are lower. Brokers adding crypto, tokenized funds and prediction markets chip at the same customer, and DBS noted prediction markets doubled quarter over quarter. Coinbase also owns Base, its own blockchain, which places it on both sides of any migration of trading onto chains.
The sharpest adverse fact in the record is cyclical. Transaction revenue was about $130 million in 3Q26 through late July, against $599 million in the whole of 2Q26, DBS reported; the periods are not comparable and part of the gap reflects a weaker market rather than lost share. The trading line can fall by more than half on crypto prices alone, which is why the moat question turns on the recurring half. DBS estimates 88% of revenue no longer comes from bitcoin spot trading, so the diversification is real even if it was partly forced.
Brand, balance sheet and engineering speed are not barriers. Coinbase is well known to US retail, holds $8.6 billion of cash, and DBS credits AI tools with a 2.2x gain in pull requests per engineer. None of that prevents a funded rival from renting liquidity or copying a product. Customer captivity reinforced by scale is the strongest combination available; Coinbase has partial versions of each, held together by a license regime and a partnership it does not control. On current evidence the advantage should survive the next cycle if the recurring share keeps rising and the USDC split holds at renewal.
Generated by DeepVal AI. Data and models may contain errors or be outdated. For research only; not personalized investment advice.
Continue reading the full report
- Comparison and summaryHow Coinbase Global Inc compares with its closest rivals, and our moat rating
Free to sign up · 20 research points · unlocking this report uses 10
20 points cover 2 full reports. No card needed.
Research reflects its generation date. Preview the first two modules.
More on Coinbase Global Inc
Coinbase Global Inc
Business, moat and key questions
Open company page