# Boomtown: The State of x402 Demand

Next: market note no. 2, The Assay: Who Actually Pays x402 Sellers, https://api.provenpaid.com/notes/who-actually-pays

October 2026. A ProvenPaid market note by Justin Smith, Ph.D. Revised 2026-10-09 to add ecosystem-wide figures from x402scan.

The x402 gold rush is real, and the miners have arrived. But the town has more shops than miners, and a handful of big outfits do most of the digging. The tools that help an agent decide where to spend are still scarce.

That is the picture from public data. Five findings:

- **The miners are here.** x402scan counts about 34,850 paying wallets and $939,000 in volume across the ecosystem in the past 30 days.
- **There are more shops than miners.** The same count shows about 40,000 sellers, and only 23 of the 202 projects in the x402.org directory build tools that help an agent pay.
- **A few big outfits do the digging.** Across the x402 sellers we measured directly, 3 wallets made 67% of 80,587 payments in 30 days. x402scan's busiest service shows the same shape: about 420,000 transactions from 149 buyers.
- **Surveyors arrive before customers.** Automated monitors found our new endpoint within minutes of launch. None of them paid.
- **The assay offices are open.** Checking that an endpoint is up is already free. Evidence of who actually pays an endpoint is still rare.

## The miners are here

x402scan, an explorer that tracks x402 payments across facilitators and chains, reports these totals for the past 30 days (read 2026-10-09, https://www.x402scan.com):

- **About 2.67 million transactions** and **about $939,000** in volume.
- **About 34,850 buyers** and **about 40,000 sellers**, by its count.
- Its busiest featured service, claw402, shows about 420,000 transactions from 149 buyers, worth about $944.

So this is not an empty town. Tens of thousands of wallets paid for something last month. Two things stand out, though. By x402scan's count there are more sellers than buyers. And the average transaction is about 35 cents, while the busiest services run on fractions of a cent repeated hundreds of thousands of times by a small number of wallets. That is what large automated buyers look like, not a crowd of agents shopping around.

These are x402scan's figures and definitions, not ours. Its methodology for counting buyers and sellers is its own, and it covers more chains than our scan.

## The town is built for shopkeepers

The folk wisdom of the California gold rush is that the merchants who sold picks and shovels did better than most of the miners. In x402, the merchants arrived early and in force.

![Bar chart: of 202 x402 directory projects, 75 are services, 68 infrastructure and tooling, 32 facilitators, 23 client-side tools that help an agent pay, 3 learning resources and 1 application.](/brand/notes/directory-by-category.svg)

Source: x402.org ecosystem directory (coinbase/x402, partners-data), 202 listings, read 2026-10-09. Categories are self-chosen.

Seventy-five services sell something. A hundred more projects are roads and banks: tooling, gateways and 32 facilitators that settle payments. Only 23 build for the buyer side, the wallets and clients an agent uses to decide what to pay for. A few wallets list themselves as infrastructure, so the buyer side is somewhat larger than its label. The shape still holds.

It is also hard to find the shop door. We sent one unpaid request to each of the 202 listed websites and to any published `/.well-known/x402` manifest. Only 6 listings led us to a live x402 endpoint that takes USDC on Base.

## A few big outfits do the digging

To see the shape of demand up close, we read 30 days of public Base payments into the receiving addresses those 6 listings advertise. Four of them received x402-style payments: 80,587 payments worth $3,892, from 394 distinct wallets. The median payment was under half a cent.

![Bar chart: the top 3 wallets made 67% of payments, the next 7 made 18%, and the other 384 wallets made 15%.](/brand/notes/payer-concentration.svg)

Source: ProvenPaid scan of public Base data, Base blocks 51,066,236 to 52,362,236 (about 30 days to 2026-10-09).

This is not a crowd of prospectors panning every stream. It looks like a few large outfits working the same claims around the clock, which matches the ecosystem-wide picture above. About 40 wallets pay on a typical day, making 2,000 to 5,000 payments between them. Only 31 of the 394 wallets, 8%, paid more than one of these sellers.

Four sellers is a small sample. Read it as a core sample from one hillside, not a survey of the territory. The territory-wide numbers are x402scan's.

## Surveyors arrive before customers

Within minutes of ProvenPaid's launch on Base mainnet, at least five automated monitors requested our paid endpoint without paying. Their user agents named themselves x402watch, x402-observer, x402-reliability-probe, agentrankprobe and carbonmonitor. Another client asked for `/.well-known/x402`.

The surveyors, mapmakers and land agents of this rush are fast and well organized. Being found is cheap. Being paid is the hard part.

## Count transactions carefully

Base settles about 80,000 signed USDC payments a day using EIP-3009, the same signature method x402 uses. Wallets and payment apps use it too, and public logs alone cannot tell them apart. x402scan, which tracks x402 activity specifically, counts about 89,000 transactions a day across all chains. The two figures are close enough that x402 may be a large share of that Base traffic, or a small one. Without a source that separates them, nobody can say. Ask how any x402 market figure tells x402 payments apart from the rest.

And count wallets, not transactions. A few hundred thousand sub-cent payments from 149 wallets is a busy pipeline, not a busy market.

## The assay offices are open

In a gold town, the assay office tested whether ore was real before anyone paid for it. x402 has its assay offices already. By our reading, about 15 of the 202 directory projects offer trust, monitoring or reputation for endpoints, payers or payments. Several check whether an endpoint is alive for free:

- **zauth** continuously probes schema, price and delivery, and alerts providers (https://zauthx402.com).
- **EntRoute** checks for a valid 402 every 10 minutes, makes paid delivery checks on top endpoints, and ranks by reliability, speed and price. Querying it is free (https://entroute.com).
- **MerchantGuard** offers a free 0 to 100 merchant and agent risk score (https://merchantguard.ai).

None of the pages we reviewed reports who actually pays an endpoint. A probe shows the shop door opens. It cannot show whether anyone other than the owner has ever bought anything.

We built one of these assay offices too. ProvenPaid reads public settlement data to report how many distinct wallets pay an endpoint, and whether they come back. The honest version of our position: most of today's spending comes from big outfits that already know which claims they work. They rarely need an assay office. The buyer who needs one is the agent trying a seller for the first time, and that buyer is still uncommon. We would rather say that plainly than pretend otherwise.

## What it means

- **For agent builders:** before your agent pays an endpoint it has not used, check two things. Does the endpoint answer, which is free? And do other wallets pay it and come back? The first guards against a dead shop. Only the second guards against a shop nobody uses.
- **For sellers:** monitors and directories will find you within minutes. That gets you on the map, not into anyone's budget. Distinct, returning payers are the evidence a careful buyer will look for.
- **For operators and investors:** the traffic is real but concentrated, and the town is crowded with sellers. Watch one number to tell when the rush arrives: the share of paying wallets that pay more than one seller. Today, among the sellers we could reach, it is 8%. When agents start shopping around, that number moves first.

## Method, limits and disclosure

- **Disclosure:** the author runs ProvenPaid, which sells payer-depth checks for x402 endpoints. Methodology: https://api.provenpaid.com/methodology
- **Ecosystem totals:** from x402scan's public home page, read 2026-10-09, using its own definitions. We did not verify them independently.
- **Directory:** all 202 listings in the public x402.org ecosystem directory, read 2026-10-09. Categories are self-assigned and descriptions unverified. The count of about 15 trust projects is our own hand classification.
- **Payment scan:** USDC `Transfer` logs into receiving addresses advertised in live x402 payment challenges, kept only when the same transaction carries the payer's EIP-3009 `AuthorizationUsed` event. Base mainnet, about 30 days to 2026-10-09, read through a commercial node provider, complete coverage. Our own test wallet and receiving address are excluded. A receiving address can serve several endpoints.
- **Reach:** 279 unpaid requests to the 202 listings and their manifests, honouring robots.txt. 6 listings exposed a Base USDC x402 endpoint, and 4 receiving addresses had payments. Sellers we could not reach this way are not counted.
- **Monitors:** user agents are self-declared and can be spoofed. We counted names, not operators.
- **Competitors:** summarized from their public pages on 2026-10-08. A feature missing from a page may still exist. Rencom was not reviewed because its site blocks automated reading.
- **Not used:** Coinbase Bazaar discovery data, under the CDP terms of service.