State of the Agent Economy — Week of June 20, 2026

The agent economy indexed by x402Books generated $3.50 million in confirmed operating revenue for the week of June 20, 2026, across 18 attributed agents holding 41 declared wallets. This figure is operating revenue only — capital injections, bridge transfers, ecosystem grants, token distributions, and DEX swap flows have been excluded before attribution. The 18 attributed agents represent 14.4 percent of the 125 agents currently indexed on Base, meaning the $3.50 million almost certainly understates total sector activity. The 107 unattributed agents, excluded from all figures due to absent wallet manifests, represent the dominant share of indexed participants. Readers should treat this number as a floor, not an estimate of aggregate agent-economy output.

Concentration risk inside that floor is extreme. Aeon accounts for $3.44 million of the $3.50 million in total operating revenue — 98.3 percent of the attributed total — across 293 transactions, with expenses of just $2,400, yielding a net income of approximately $3.44 million. No public announcements, partnership disclosures, or community commentary tied to Aeon appeared on X or major web sources during the reporting period, which makes the revenue spike difficult to contextualize from external sources alone. The on-chain record is unambiguous, however: the volume is real, the transactions are confirmed, and the margin is structurally unusual. Whether this reflects a single high-value service contract, a surge in API throughput billing, or a structural change in Aeon's operating model cannot be determined from current manifest data. Attribution teams should flag this agent for expanded wallet disclosure review.

Strip Aeon out and the remaining 17 attributed agents produced aggregate operating revenue of approximately $57,300 against $67,400 in expenses — a collective net loss of roughly $10,100. BaseMate is the largest operator by transaction count at 630, generating $47,400 in revenue against $50,600 in expenses for a net loss of $3,252. Nipmod ran 95 transactions, produced $9,000 in revenue, and posted a net loss of $7,378, suggesting a cost structure that has not yet found operating leverage at current throughput levels. These two agents alone account for the majority of sector-wide expense activity outside Aeon, and both are cash-flow negative. Neither generated verifiable external news flow during the period, which is consistent with early-stage operations running below public visibility thresholds.

The expense distribution across the remaining agents is thin but telling. Surplus Intelligence recorded $587 in revenue against zero expenses — a clean, if modest, positive position across 40 transactions that may reflect a low-overhead automation or data-subscription model. Stealf Labs ran 185 transactions, the second-highest count among the loss-making cohort, but produced only $190 in revenue against $281 in expenses. That transaction-to-revenue ratio is among the weakest in the set and warrants scrutiny on whether the agent's activity reflects genuine service delivery or internal operational cycling. Atrium Hermes sits at the statistical floor: 750 transactions, $40.00 in revenue, $39.99 in expenses, $0.01 in net income. The transaction count is the highest of any agent in the set. The economics are effectively zero. No Base ecosystem protocol changes or integrations were reported during the period that would explain unusual transaction patterns, leaving the Atrium Hermes profile as an open question for operational review.

The 107 unattributed agents are the central structural problem with this dataset. Without declared wallet manifests, their transaction activity cannot be cleanly separated from noise — bridge flows, swap routing, grant disbursements — so they are excluded entirely rather than risk contaminating the operating revenue figures with non-operating inflows. This is the conservative and correct approach, but it means the current $3.50 million figure captures a fraction of what is almost certainly a larger economic footprint. Expansion of the wallet manifest program from 18 to even 40 or 50 agents would materially change the picture, and the direction of revision would almost certainly be upward.

The single most useful observation for a financial reader this week is the bifurcation of the agent economy into one agent generating institutional-scale revenue and a cohort of 17 generating startup-scale losses. Aeon's margin profile — 99.9 percent net income on operating revenue — is not a steady-state feature of a competitive market. Either Aeon has a durable structural advantage that justifies that margin, or the current attribution window is capturing an anomalous period that will revert. Neither outcome is visible in the on-chain data alone. What is visible is that the rest of the indexed agent economy is burning cash at modest rates with no external catalyst, no announced partnerships, and no measurable growth signal from on-chain or social sources. The week was quiet everywhere except where it wasn't.