Price
$163.41
−3.2% on the day
P / FCF (TTM)
~16x*
*reported — see §02 caveat
52-wk range
$139–445
near the low (−63% off high)
USDC revenue
$1.35B
+48% YoY · 19% of rev
The one-page take
Coinbase is two businesses wearing one ticker. ~58% of revenue is trading — cyclical, sentiment-driven, and currently shrinking. But underneath it, a recurring "subscription & services" engine is compounding, led by stablecoin (USDC) revenue at $1.35B, growing ~48%/yr — now 19% of the company and rising.
The asymmetry: COIN co-created the x402 agent-payments standard, owns Base (the chain most x402 volume settles on), and earns a high-margin revenue-share on every USDC dollar — much of it at near-zero incremental cost. So it's a profitable, cash-generative, capital-returning business (it started buying back stock) that also holds a free option on machine-to-machine payments.
The catch: you're buying crypto-cycle beta, and Coinbase's GAAP metrics lie. Net income is dominated by mark-to-market on its own crypto (TTM operating income is ~$0; the "profit" is a crypto-price bet), and reported FCF is flattered by customer-custodial timing — so the ~52x P/E and ~16x P/FCF are both misleading. Value it on adjusted earnings, and the real debate is cheap cyclical vs. value trap at a cycle peak. See §02.
01What you actually own — revenue mix
FY2025 revenue ≈ $7.2B. The headline "crypto exchange" hides a fast-growing recurring layer.
Trading 58%
USDC 19%
Staking 10%
Other 13%
Transaction / trading (consumer + institutional)
Stablecoin / USDC revenue-share
Blockchain rewards (staking)
Other subscription & services
| Segment ($B) | 2023 | 2024 | 2025 | '24→'25 |
|---|
| Consumer trading (net) | 1.43 | 3.43 | 3.32 | −3% |
| Institutional + other trading | 0.09 | 0.56 | 0.73 | +30% |
| Stablecoin (USDC) | 0.69 | 0.91 | 1.35 | +48% |
| Blockchain rewards (staking) | 0.33 | 0.71 | 0.68 | −4% |
| Other subscription & services | 0.21 | 0.42 | 0.55 | +31% |
| Other revenue | 0.18 | 0.27 | 0.30 | +11% |
Why the mix matters
Trading is ~58% of revenue and flat-to-down — that's the cyclical, hard-to-forecast half. The bull case lives in the other half: USDC revenue has gone $0.69B → $0.91B → $1.35B, the single fastest-growing major line, and it's recurring and rate-linked rather than volume-linked. As USDC scales, this line has real operating leverage.
02Valuation
| Metric | Reported | What it really is |
|---|
| Price / Market cap | $163.41 / ~$43B | — |
| P / E (TTM) | ~52x | noise — see below |
| TTM net income | $0.80B | ~all crypto MTM |
| TTM operating income | ~$0.03B | ≈ breakeven |
| P / FCF — reported | ~16x | flattered (see ladder) |
| P / FCF — ex-custodial-timing | ~23x | strips WC noise |
| P / FCF — owner earnings (ex-SBC) | ~46x | the honest floor |
| P / Sales · P / Book | 7.7x · 3.3x | cleaner top-line gauges |
Bottom line on the multiple: COIN is not cheap
The ~16x P/FCF is a mirage; 23x flatters it (treats ~$0.9B/yr of stock comp as free); honest owner earnings is ~46x; and TTM operating income is ~$0. It's a fully-priced quality franchise, not a value stock — and violently cycle-dependent (Q1'25 operating income +$706M vs Q4'25 −$444M). The "16x" caught a favorable trailing window that's already rolling over. Even the street's most COIN-flattering lens — adjusted EBITDA (adds back both crypto-MTM and SBC) — is only ~mid-teens EV/EBITDA. You're paying full price for a better business; you are not getting it cheap.
Read the multiples carefully — Coinbase's GAAP is uniquely distorted
- P/E is meaningless. Since accounting rule ASU 2023-08 (2025), Coinbase marks its own crypto holdings to fair value through net income. TTM operating income is ~$0; the $0.8B "profit" is a non-operating crypto-price bet. Q2'25 showed +$1.43B net income on −$25M operating income (a +$1.85B crypto gain); Q4'25 and Q1'26 were GAAP losses as it reversed.
- P/FCF is flattered. Reported TTM FCF ($2.79B) includes ~$0.9B of customer-custodial-cash timing (one quarter, Q4'25, had +$3.06B OCF on a working-capital swing; Q3'25 was −$785M). Normalized owner FCF is ~$1.8–1.9B → ~23–24x, richer once you charge ~$0.9B/yr of stock-based comp.
- Value it on adjusted EBITDA / adjusted net income (ex-crypto-MTM, ex-SBC) — the way the sell-side does — not GAAP P/E or reported FCF.
03Cash flow & balance sheet
The quality argument: real, growing free cash flow — and it's now being returned to shareholders.
| $B | 2022 | 2023 | 2024 | 2025 |
|---|
| Revenue | 3.19 | 3.11 | 6.56 | 7.18 |
| Operating income | −2.71 | −0.16 | 2.31 | 1.44 |
| Net income | −2.62 | 0.09 | 2.58 | 1.26 |
| Free cash flow | −1.59 | 0.92 | 2.56 | 2.43 |
| Buybacks | — | — | — | −0.79 |
Balance sheet: cash ~$11.3B, total debt ~$7.8B (net cash ~$3.5B), equity ~$14.8B. Strong, self-funding, and — unlike Circle — returning capital rather than diluting.
Caveat (see §02): annual FCF here is real but lumpy — quarterly OCF swings ±$1B on customer-custodial-cash timing, so treat any single-period FCF with care and normalize to ~$1.8–2B.
The cyclicality, in one row
Look at 2022: revenue halved, FCF −$1.59B, net loss −$2.6B. That is the drawdown you are underwriting. Trading revenue is levered to crypto prices and volumes; a bear market hits COIN's biggest line directly. This is the central risk — not USDC, not x402.
04The USDC / stablecoin engine
USDC is issued by Circle (CRCL), but Coinbase is its distribution partner — and the economics are unusually favorable to COIN:
- Coinbase earns 100% of the reserve income on USDC held on the Coinbase platform, and roughly 50% of the residual reserve income on USDC held elsewhere.
- That revenue-share arrives at near-zero incremental cost — it drops almost straight to margin.
- Result: COIN's stablecoin line is $1.35B and +48% YoY, while Circle — the actual issuer — keeps only ~19% of gross reserve income after paying Coinbase and others.
Market context
USDC circulation is ~$73B today (2026-07-06), the #2 stablecoin, live on 30+ chains and expanding fast (heavy 2026 minting on Solana). COIN's stablecoin revenue is a direct function of (circulation × short-term rates) — so it's a rate play and an adoption play, captured at a better margin than owning Circle itself.
05The x402 / agentic-payments option
This is the part that isn't in the numbers yet. "Agents with wallets" (e.g. Franklin/BlockRun) autonomously spend USDC per-action via x402, an open HTTP-native micropayment standard:
- Coinbase co-created x402 (with Cloudflare) — now a Linux Foundation project (formalized Apr 2026) with 22+ members incl. Visa, Mastercard, Stripe, Google, AWS, Circle, Shopify.
- Coinbase owns Base, the L2 where most x402 volume settles, and runs the Agent.market directory of paid agent services.
- Every x402 transaction is a USDC settlement — feeding the same revenue-share engine above.
Size the option honestly
As of late-April 2026, all of x402 was ~69K active agents, 165M transactions, and ~$50M cumulative volume — trivial vs. Visa's ~$16T/yr. This is a free option, not a revenue line. The single metric to track: x402 cumulative volume going from tens-of-millions to billions. Until then, value COIN on the trading + USDC businesses and treat agentic payments as upside.
06Scenario — what if USDC circulation goes to $3T?
A stress-test of the bull case. $3T is ~41x today's ~$73B — a "thesis-wins-spectacularly" number, not a base case.
Circle's own math: at $3T, gross reserve income is enormous ($60–120B depending on rates), but the issuer keeps <20%. Coinbase sits on the other side of that split — it collects its revenue-share on the same $3T at near-zero incremental cost. So a USDC explosion flows heavily into COIN's stablecoin line, arguably at a higher incremental margin than into Circle's P&L.
| If USDC = $3T | Reserve yield 4% | 3% | 2% |
|---|
| Gross reserve income (whole USDC pool) | $120B | $90B | $60B |
| Circle keeps (~15–19%) | ~$22B FCF | ~$14B | ~$8B |
| Coinbase share (high-margin skim) | a large, low-cost multiple of today's $1.35B line |
Takeaway: even Circle — the pure-play — likely does "only" a 5–30x on a 41x circulation move (rate risk, take-rate compression, multiple compression). COIN captures a big share of the same upside with far less single-variable risk, because USDC is a growing kicker rather than 100% of the story.
07COIN vs. CRCL — the head-to-head
| COIN | CRCL (Circle) |
|---|
| Market cap | ~$43B | $18.3B |
| P / FCF (reported) | ~16x* | ~35x |
| FCF yield | ~6.3%* | 2.8% |
| 2025 FCF | $2.43B* | $0.53B |
| GAAP net income | +$1.26B FY25
(TTM ~$0.8B, crypto-MTM) | −$70M (IPO SBC) |
| Capital return | buyback | dilutes (SBC) |
| USDC as % of revenue | ~19% (kicker) | ~100% (pure-play) |
| Dominant risk | crypto cycle | rates + Coinbase take |
| Owns the rail? | Base + x402 | the currency (USDC) |
* COIN's reported FCF/P-E are flattered by custodial-cash timing and crypto mark-to-market (§02); normalized P/FCF is ~23–24x. The comparison still favors COIN, just by less than the raw ~16x-vs-35x implies.
Verdict
Better business + cheaper stock → COIN. Half the P/FCF, ~4.6x the free cash flow at ~2.4x the market cap, profitable, returns capital, and it owns the rail while still capturing USDC's high-margin skim. Higher-torque pure-play on "USDC circulation explodes" → CRCL. If you believe the whole thesis, own both; the one thing to avoid is buying CRCL instead of COIN thinking it's the safer stablecoin play — it's the riskier, more concentrated, more expensive one.
08Positioning — short interest
| COIN | CRCL |
|---|
| Shares short | 26.8M | 22.3M |
| % of float | 12.2% | 12.4% |
| Days to cover | 3.0 | 1.28 |
| Trend vs. prior | +9.85% | — |
Both ~12% short. COIN's short is more crowded relative to liquidity (3.0 days to cover vs 1.28) and rising — a modestly more bearish tilt building on COIN, and a mild squeeze setup if the crypto tape turns. CRCL's 1.28 days = extremely liquid, low squeeze risk despite the 12%.
09Open questions for the research
- Rate sensitivity of the USDC line. How much of the $1.35B stablecoin revenue is pure reserve interest vs. other? What does a 100bp Fed cut do to it, and how much is offset by circulation growth?
- Normalized trading economics. Through a full cycle, what is COIN's mid-cycle trading volume and take rate? What does a bear-case trading year look like on FCF?
- Why did 2025 margins compress? Operating income fell ~38% ($2.31B → $1.44B) on higher revenue — S&M, international expansion, tech spend, or take-rate? Is it structural or investment?
- Base & x402 monetization. How does Coinbase monetize Base today (sequencer fees, etc.), and what's the roadmap to turn agentic-payment volume into disclosed revenue?
- Regulation. How do stablecoin rules (GENIUS Act and successors) affect the Circle↔Coinbase revenue-share economics — helpful, neutral, or a renegotiation risk?
- Circle relationship durability. The revenue-share is contractual — when does it renew, and what leverage does each side have?
- Capital return trajectory. Is the 2025 buyback a one-off or the start of a program? What's the FCF-payout intent?
- The core valuation question. Is ~16x FCF a cheap entry on a compounding rails business, or a value trap set at a crypto-cycle peak? What's the right normalized-FCF multiple?