48 Networks, One Winner: What 24,419 x402 Resources Actually Settle On
"Which chain does this x402 service settle on?" is a question the catalog can answer and the homepage copy usually can't. So we paged the entire CDP x402 discovery index โ the largest public x402 catalog โ and counted what its resources actually advertise as payment options. The result is cleaner than the marketing: 48 distinct networks appear, and Base appears in 97.9% of resources. The interesting part is not that one chain dominates. It is how much of the catalog is one publisher, and how fast it grew without diversifying.
What was measured
- Source:
https://api.cdp.coinbase.com/platform/v2/x402/discovery/resourcesโ the CDP x402 discovery index, read-only, no auth. - Method: paginate the endpoint, combine every page, count the
accepts[]array on each resource. Each entry inaccepts[]is one payment option: a network, an asset, an amount, a payee. - Frame: 24,419 resources on 2026-10-03. The same index read 19,126 resources on 2026-09-29 โ up 27.7% in four days.
accepts[] entry is an advertised
option. Neither is evidence that anyone paid. This is a map of what the catalog says is payable,
not of what settled.
Finding 1 โ Base is present in 97.9% of resources
Grouping by whether a resource's advertised networks include Base (eip155:8453):
| Class | Resources | Share |
|---|---|---|
| Base only | 14,515 | 59.4% |
| Base + at least one other network | 9,398 | 38.5% |
| No Base at all | 506 | 2.1% |
Read together with the asset column below, this says something practical for a buyer agent: if it holds USDC on Base, it can pay 97.9% of the advertised catalog. If it holds USDC on any other single chain, it can pay a small fraction of it. The catalog's diversity lives almost entirely in the multi-rail tail, not in the base case.
Finding 2 โ 48 networks, and the tail is long and thin
| Network (CAIP-2) | Resources advertising it |
|---|---|
eip155:8453 (Base) | 23,913 |
solana:5eykt4UsFv8P8NJdTREpY1vzqKqZKvdp | 6,638 |
eip155:137 (Polygon) | 3,408 |
eip155:42161 (Arbitrum) | 2,665 |
eip155:84532 (Base Sepolia) | 2,366 |
eip155:143 | 910 |
eip155:10 (Optimism) | 659 |
eip155:43114 (Avalanche) | 655 |
xrpl:0 | 641 |
eip155:42220 (Celo) | 603 |
stellar:pubnet | 599 |
algorand:wGHE2Pwdvd7S12BL5FaOP20EGYesN73ktiC1qzkkit8= | 337 |
The 36 networks below the visible cut are real but small. Note that Base Sepolia โ a testnet โ carries more than a quarter as many listings as Solana mainnet. A discovery index counts what people list, not what people pay, and testnets list cheaply.
Finding 3 โ the asset is USDC, and the scheme is exact
Across every accepts[] entry in the catalog, the asset is overwhelmingly USD Coin
(the same symbol and standard, deployed per chain). The five most-quoted assets, by number of
entries:
| Asset (contract) | Accept entries |
|---|---|
0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913 (USDC, Base) | 25,186 |
EPjFWdd5AufqSSqeM2qN1xzybapC8G4wEGGkZwyTDt1v (USDC, Solana) | 6,893 |
0x3c499c542cEF5E3811e1192ce70d8cC03d5c3359 (USDC, Polygon) | 3,430 |
0xaf88d065e77c8cC2239327C5EDb3A432268e5831 (USDC, Arbitrum) | 2,711 |
0x036CbD53842c5426634e7929541eC2318f3dCF7e (USDC, Base Sepolia) | 2,371 |
The settlement scheme is just as concentrated. Of all accept entries, exact accounts
for 48,257; everything else is a long tail โ upto (782),
batch-settlement (52), and a dozen experimental schemes in the single and double
digits. A catalog of 24,419 resources is, in practice, a catalog of one asset and one scheme with a
decorative fringe.
Finding 4 โ 2,054 publishers, and the top of the curve is steep
Counting by resource host:
| Measure | Value |
|---|---|
| Distinct hosts in the catalog | 2,054 |
| Hosts with โฅ100 resources | 36 |
| Share of all resources held by the top 5 hosts | 28.8% |
One host alone lists 3,420 resources. That single fact changes how the whole catalog should be read: a 27.7%-in-four-days jump in the resource count is not 5,000 new teams arriving. It is a handful of bulk publishers pushing rows. The index is inflating faster than it is diversifying, and the two numbers above โ 24,419 resources, 2,054 publishers โ are the pair to quote together, never the first alone.
exact-scheme seller. That is the median of the catalog by rail, not the
tail. It is also, deliberately, a single rail โ we do not advertise Solana, XRPL, Celo, Stellar, or
Algorand, and the 2.1% of resources that take no Base are, for us, unreachable buyers in the other
direction. Cross-rail paying is the open problem this data makes visible: the catalog lists a
buyer's chain, but a 402 is answered in one rail, and nothing in the listing reconciles the
two.
What a builder should take from this
- Quote the publisher count, not just the resource count. "24,419 resources" and "2,054 publishers" are the same catalog; only one of them is a market size.
- Base is the safe default rail, by a wide margin. 97.9% of advertised resources accept it. If you ship one settlement rail, ship that one.
- Testnets inflate listings. Base Sepolia out-lists XRPL and Stellar mainnets in this index. Filter testnets out before you draw any conclusion about mainnet coverage.
- "Multi-chain" in a listing usually means "one chain plus a fallback." 38.5% of
resources are multi-network, but the currency is nearly always USDC and the scheme nearly always
exactโ the variety is in the chain, not in the payment shape.
Source: api.cdp.coinbase.com/platform/v2/x402/discovery/resources, full pagination
read on 2026-10-03 (24,419 resources). Compared against the same index read on 2026-09-29
(19,126 resources). All figures are counts over the resources' own advertised
accepts[] arrays; nothing here is a settlement figure. Re-derivable from the public
endpoint โ the numbers will move, the base/asset/scheme concentration has held across both reads.