Agent Economy Report

Monthly report

State of the Agent Economy

October 2026 · data through 2026-11-01 · updated monthly · A research initiative of the Center for Studies in Strategy and Innovation, COPPEAD Graduate School of Business, UFRJ · Em português

Machines started paying machines through the x402 protocol. Every dashboard shows the gross flow. This report shows what is left after removing storefronts and automated distributors, who the money goes to, how many services anyone actually pays, and how much settles where no catalog is looking.

Settlement, day by day

USDC settled per day to x402 services at agent scale, across Base, Polygon, Arbitrum and Solana (the Agent Economy Index). Last 7 days: $10.3k (-5% week over week). By network: base 92%, solana 8%, arbitrum 0%, polygon 0%.

Who gets paid

2,799 wallets received USDC in 30 days; 2,379 of them at agent scale. Two wallets took 76% of the gross. Among rated services, 1919 had any settlement in 14 days and the top 10 took 66% of it.

Trust

2,041 services rated by proven adoption (organic paying agents, real settlement, network centrality, uptime as a gate), on a scale from AAA to D with the trust line at BBB. Separately, 513 unlisted wallets are rated from on-chain evidence alone, capped at BB.

Above the trust line, day by day: 10-23 52 · 10-24 53 · 10-25 53 · 10-26 56 · 10-27 53 · 10-28 57 · 10-29 56 · 10-30 54 · 10-31 55 · 11-01 55.

Method change on 2026-10-03. Liveness probe by method. Before this date the daily probe sent an unpaid GET to every listed resource and counted a 405 Method Not Allowed as downtime, so a service whose resources accept only POST was rated as down. From this date a 405 to the GET is retried with POST, a 402 (or 200) to either method counts as up, and a resource that answers the POST is credited for its earlier 405 answers within the window. Uptime, and with it the tier, can rise on this date because of the rule, not because the service changed. The rule.

Method change on 2026-10-03. Sweepers are not organic. Before this date every payer that paid more than one service counted as an organic paying agent, so monitors and sweeps that pay hundreds of services in a day were counted as adoption by every one of them, and the index moved in blocks when such a day entered or left the window. From this date a payer that paid 30 or more different wallets on a single day of the window is not organic: it does not count for adoption, the trend or the centrality graph. Its payments still count as settlement, and the CAPTIVE test leaves sweepers out of its count. Adoption, and with it the tier, can move on this date because of the rule, not because the service changed. The rule.

Method change on 2026-10-22. Tier hysteresis. Before this date a tier moved the moment the score crossed a threshold, so entries sitting near one moved back and forth with the day-to-day noise of a relative score. From this date a tier changes only when the score clears the threshold by 2 points: it rises only to a tier the score still reaches 2 points lower, and falls only when the score 2 points higher would still fall. Thresholds that do not depend on the score (uptime, organic agents, age, CAPTIVE, NEW, the BB cap of unlisted wallets) still apply at once. The score itself is unchanged, and each entry also publishes the tier the day's numbers alone would give (tier_raw). No tier rises on this date because of the rule; some fall later than they would have. The rule.

Method change on 2026-10-28. Agent-scale ceiling per platform. Before this date the ceiling of 100,000 transfers in 14 days was checked wallet by wallet, so a closed platform that spreads its payments over thousands of wallets stayed under it with every one of them. From this date wallets outside the catalogs that are linked by payers who pay nothing in the catalogs are one platform when at least 90% of each wallet's payers, and the payers of at least 90% of the value it receives, are such payers and also pay one of them (the hub), and the same holds for the hub with the payers of the other wallets; a platform above the ceiling leaves the index and the unlisted watchlist, as any single wallet above it does. Below the ceiling nothing changes, and no other entry is recalculated. Unlisted entries can end on this date because of the rule, not because the service changed. The rule.

Method change on 2026-10-31. Payments with the receiver's own money. Before this date the Solana reader counted every USDC transfer into a payTo wallet as a payment by the wallet that sent it, so a service that funds wallets and has them pay it back was counted as having that many paying agents and that much settlement. From this date, on Solana, two kinds of payment do not count, for settlement, adoption, centrality, the AEI or the monthly report: those of a wallet that the receiving payTo wallet itself funded with USDC from an empty balance and that then paid it, both in the same day's reading (the wallet stays marked for that payTo wallet; one that already held USDC, one funded only after its first payment of the day, or one that had already paid it on an earlier day is not marked by this part); and the payments a wallet made to a payTo wallet on a day when that payTo wallet sent it, for any reason (a refund, a payout, a payment for a service between the two), at least half of what the wallet paid it, as read that day (that day only). Only direct transfers are detected, and only on Solana. Applied from the data of this date; earlier days are as they were read. Settlement and adoption of a service, and with them its tier, can fall on this date because of the rule, not because the service changed. The rule.

Catalogs

38,197 resources listed by 4,171 hosts across the two discovery catalogs (CDP Bazaar and PayAI) on 2026-11-01. 753 hosts appeared in the CDP Bazaar in the last 7 days. Liveness probe on 2026-11-01: 3,309 of the 4,171 hosts listed that day (79%) answered with a price (HTTP 402) on at least one of their listed endpoints. By resource, of the 38,197 resources listed that day, 85% answered with a price and 1% did not answer at all. The share by resource follows the largest hosts: saylorinnovations.com alone lists 5% of the listed resources and answered with a price on 99% of its own resources that day. The probe also covers every resource listed since August (185,093 on 2026-11-01; 30% answered).

The part nobody indexes

Every x402 facilitator settles USDC with transferWithAuthorization (EIP-3009). We read all of those authorizations on Base, every day, and keep the receivers that are in no discovery catalog. In the last 3 days, agent-scale wallets outside the catalogs settled $869 against $3,545 inside them (20% outside). Wallets of a closed platform above the agent-scale ceiling are left out, as from the index (the rule). The largest unlisted wallet, 0x47d3394c…, took $1,950 from 28 payers in 14 days and has no name anywhere. Live list: the unlisted watchlist.

Get the next one

One e-mail a week with the index, the rating actions and, once a month, this report. Unsubscribe in one click.

Off-chain: the agent labor supply

Most agent work is paid by subscription, not on-chain (Grok Bot, OpenClaw, Claude Cowork). We cannot measure that settlement, so we measure the supply: the ClawHub registry listed 81,858 skills on 2026-11-01, 105,108,768 downloads in total, 678 skills added in 7 days and 3,131 in 30; 81,476 have 100+ downloads and 862 mention x402, wallets or USDC. Most downloaded: Self-Improving + Proactive Agent (213,599), Github (203,532), ontology (201,082), SkillScan (183,108), Proactive Agent (178,755).

Method

Every figure is computed from the Agent Economy Report daily archive: the x402 discovery catalogs (CDP Bazaar and PayAI), USDC transfers to every catalog payTo wallet on Base, Polygon, Arbitrum and Solana mainnets, EIP-3009 AuthorizationUsed events on Base (all facilitators), and a daily liveness probe of every listed resource. On Solana, from 31 October 2026, payments with the receiver's own money are left out: those of a wallet that the receiving payTo wallet itself funded from an empty balance earlier in the same day's reading (and its later payments to that wallet), and a wallet's payments on a day when the payTo wallet sent it at least half of what it paid. Agent scale = payments read per payer, wallet and day: a day averaging at most $1 counts in full, on a day averaging above $1 each payment counts as $1, dust under $0.0001 is not counted; wallets need at least one payment of at most $1, no fewer of those than payments above $1 and <= 100,000 transfers other than dust in the trailing 14 days. The rest (the part of larger payments above the cap, dust, human storefronts, automated distributors) is reported as gross but excluded from the index. Ratings: sybil-resistant adoption score, trust line at BBB.

CC BY 4.0. Cite as: Center for Studies in Strategy and Innovation, COPPEAD Graduate School of Business, UFRJ (2026). State of the Agent Economy, October 2026. Agent Economy Report. https://agenteconomy.report/report/2026-10. Machine readable: 2026-10.json. Daily: aei.json, ratings.json, unlisted.json.