x402 market
Weekly numbers on x402, the protocol AI agents use to pay for things, after removing the payments that aren’t real demand. Every chart says what it is out of and has its data underneath.
Data through 2026-09-27 · updates Mondays
55% of settled x402 dollars survive cleaning; sellers with a single buyer remove the most ($59k)
USD settled on Base in the week of 21 Sep 2026 (593,454 payments), and what each filter removes, in the order the index applies them. Manufactured volume (loops, shared funding, fan-out) is $2,505 (1.9%).
Dropping single-buyer sellers moves the headline most: without that filter, clean volume would be $132k (+80%)
Demand-cleaned USD in the week of 21 Sep if sellers needed at least 1, 2, 3 or 5 distinct genuine buyers; every other filter unchanged. The published index uses 2.
| Min. genuine buyers per seller | Clean USD | vs published | Sellers | |
|---|---|---|---|---|
| 1+ | $132k | +79.9% | 27,706 | |
| 2+ (published) | $73k | – | 637 | |
| 3+ | $73k | −0.4% | 407 | |
| 5+ | $70k | −4.2% | 213 |
Large ticket (unidentified) is 51% of demand-cleaned spend; publisher-style content is $44.66 (0.06%)
Demand-cleaned USD (or payments) per week by what the seller sells, 24 Aug – 21 Sep (5 weeks). Categories come from sellers' own listings; "unclassed" sellers have none.
The top 10 sellers take 94% of demand-cleaned spend; in compute & tools, one seller takes 98%
Share of demand-cleaned USD earned by the largest seller and the 10 largest, per week. Week of 21 Sep: 637 sellers, HHI 2,306, 4.3 effective sellers.
8,176 genuine buyers paid in the week of 21 Sep; 53% were new
Distinct payer wallets with a demand-cleaned payment, split by whether they paid the week before. The first week has no history. Median week-1 retention of new buyers: 22%.
93% of genuine buyers pay only one seller
Genuine buyers in the week of 21 Sep by how many distinct sellers they paid (multi-homing). Denominator: 8,176 buyers.
The median genuine buyer spent $0.0745 in the week
Genuine buyers by total demand-cleaned spend in the week of 21 Sep, log bins. Denominator: 8,176 buyers. Buyers page →
The median demand-cleaned payment is $0.003; large ticket (unidentified) runs 3597× that
Payment size per category in the week of 21 Sep: bar = p10 to p90, dot = median. Denominator: 537,333 demand-cleaned payments; categories with under 20 payments are hidden.
Coinbase settles 99% of demand-cleaned dollars
Share of all settled USD (grey) vs demand-cleaned USD (gold) by facilitator, week of 21 Sep. Addresses grouped by facilitator name. Denominators: $134k raw, $73k clean.
Content + data: $1,914 this week, −20% over 4 weeks — below the +20%/month benchmark
Demand-cleaned USD to content and data sellers per week vs the path of +20% a month from the first week. Log-linear trend over the last 5 weeks: −51% a month. The tripwire needs +20%/month for 3 months running.
Movers
23 sellers crossed 5 genuine buyers in the week of 21 Sep and 110 dropped below it. Biggest gain: 0xffde1acb… ($27k); biggest loss: 0xed9fcd0d… (−$2,612). Only sellers with at least 5 genuine buyers in either week are listed, so a seller whose only customer is itself does not count.