"Why not just use Stripe?" is the first question any x402 pitch gets, and it deserves a real answer instead of a protocol-purity one. Stripe is mature, trusted, and handles subscriptions, refunds, disputes, and compliance better than a wallet-and-a-signature ever will. x402 doesn't compete with that. It exists for the price range Stripe's own fee structure makes structurally uneconomical — and the size of that range is bigger than most people building agent tooling in 2026 have priced out.
The number that actually decides this: $0.30
Stripe's standard US card processing fee is 2.9% + $0.30 per successful charge — published, stable, not a rate anyone negotiates down at low volume. Run that against a $0.01 API call: 2.9% of a cent plus the fixed $0.30 is a $0.3003 fee on a $0.01 transaction — over 3,000% of the transaction's value. There's no version of that pricing that works. You either bundle micro-calls into a subscription (killing the pay-per-use model), pad each call's price until the fee looks reasonable (killing the "cheap enough to call per-request" pitch), or don't charge for that tier of usage at all.
That's not a Stripe flaw — a fixed per-transaction fee makes sense for a system built around $5-$5,000 charges with real fraud/dispute risk to underwrite. It's just a fee structure with a floor, and the floor sits well above what a lot of agent-facing API calls are actually worth.
What real x402 sellers are actually charging
Fetchgate's own x402 Services & Facilitator Registry — 118 live x402 services, collected 2026-08-21 — has real pricing for 53 of them (the rest price per-tool or don't publish a flat rate). Here's the distribution:
| Lowest priced service | $0.001/call |
| Median priced service | $0.01/call |
| Highest priced service | $50 (a data-heavy outlier, not typical) |
| Priced under $0.01/call | 22 of 53 (41.5%) |
| Priced under Stripe's $0.30 fixed fee | 45 of 53 (84.9%) |
Read that last row again: 85% of real, live x402 services in this dataset are priced below what Stripe alone would charge to process one transaction. That's not a hypothetical price point some spec author picked — it's what sellers are actually charging today, for things like web scraping ($0.001-$0.008 per call, per-tool), search, on-chain data lookups, and inference. At that price, a payment rail with no fixed per-transaction floor isn't a nice-to-have; it's the only way the pricing works at all.
Where the two actually overlap
The comparison isn't clean-cut, and pretending it is would undersell both tools:
- Subscriptions, one-time purchases $1+, anything needing chargebacks/disputes: Stripe, clearly. x402 has no dispute mechanism — a settled payment is final, which is a feature for a $0.002 API call and a serious liability for a $200 purchase.
- Sub-$0.30 pay-per-call, especially agent-to-agent with no human in the loop: x402. No account, no API key provisioning, no PCI scope, no minimum viable transaction size.
- Human-facing checkout, even for a cheap digital product: still usually Stripe or Gumroad-style processors — most humans have a card, not a funded crypto wallet, so a x402-only checkout with no fallback locks out your actual buyers. (Fetchgate's own storefront runs both: x402 for agents, a walletless Gumroad checkout for humans, same catalog.)
- High-frequency, low-value, machine-initiated calls (the exact shape of an agent calling a tool a few thousand times a day): x402's per-call settlement without a $0.30 floor is the only economically sane option once you're below roughly a dollar.
Chain and category reality check
Of the 118 services in the registry, 87 settle on Base, 14 on Solana, and a handful on Avalanche/Arbitrum/Polygon — Base's dominance mirrors the facilitator landscape (see the facilitator comparison), and 96 of 118 (81.4%) were independently verified live at collection time, not just listed. The category spread — Data (24), AI (10), On-chain Data (7), Finance (7), Web Search (5) — is a reasonable proxy for where agent-facing pay-per-call demand actually concentrates right now: structured data and inference calls that are individually cheap but too numerous to provision API keys for one at a time.
The honest verdict
If your product needs subscriptions, refunds, or human checkout, x402 isn't replacing Stripe for you — it's not trying to. If you're pricing individual API calls or tool invocations under roughly $0.30 and your buyer is another agent rather than a human with a credit card, Stripe's fee floor isn't a rounding error at that price point, it's the whole problem, and x402 is the answer that actually exists today (not a future roadmap item) for that specific gap.
This is a snapshot of real pricing data from the x402 Services & Facilitator Registry ($15) — 118 services, 21 facilitators, chains, fee models, and verification status per entry — see a free sample of 10 services first. If you're building the payment side yourself — seller or buyer — the x402 Quickstart Kit ($7) has working reference code in TypeScript and Python, including the exact facilitator wire format both need to speak.