Our census measured the supply side, what exists and how it behaves. This is the demand side, who is actually paying and how much. It is harder to measure honestly, so every number here is a bound, and we say which.
As of 2026-07-22 00:46 UTC · on-chain figures are trailing 30 days on Base · every number is a bound.
The same ecosystem looks enormous, modest, or negligible depending on what you count. Transactions, raw dollars, and price-matched dollars give three very different answers, and only the last is close to real demand. We show all three, and every figure here is a bound, not a fact.
Over 30 days we index about 13,299,954 USDC transfers to service wallets on Base. One payer-and-receiver pair is 95% of them: one client calling one routing endpoint several times a second. But it dominates only the count. By value that pair is about 17% of the USDC, because each payment is a fraction of a cent. A payment count tells you nothing about demand.
Set the routing pair aside and about $760,522 in USDC reached service wallets over 30 days. But that is inflow from every source, not x402 payments, and about two thirds of it lands on a single commerce endpoint whose wallet also collects its ordinary, non-x402 business. Wallet inflow is a ceiling, not a measure of what a service earns through x402.
The honest floor is USDC that arrives in an amount matching a service's advertised price. Across all 1,529 rated services that totals about $36,505 over 30 days. 87% of it is a single commerce endpoint, the top ten services are 94%, and only 425 services show any price-matched settlement at all. Remove that one endpoint and the entire rest of the price-matched x402 economy is about $4,824 over 30 days.
This is a floor. Exact-amount matching is blind to dynamically priced services, off-chain settlement, and other chains, so true demand is higher. The point is not the exact figure; it is that the verifiable, price-matched economy is small and concentrated in one commerce use case.
The closest proxy for a working agent is a wallet that pays several different services, repeatedly. About 1,681 wallets do, carrying roughly $47,958. But the most active of them are not customers: a few pay hundreds of different services across every category in tiny, uniform amounts, the signature of automated coverage or testing rather than demand. Strip those out and genuine multi-service use is a thin band of wallets and a few thousand dollars. There is no large agent economy settling on-chain yet.
By the money that can be verified on-chain, x402 today is real but early: supply is being built well ahead of demand. That is the normal shape of new infrastructure, not a criticism, and it is exactly why an independent measure matters now, before the narrative outruns the numbers. We will keep measuring, and when demand arrives this page will show it.
See the supply-side companion, the state of x402, measured, and the live register at x402register.com.