State of Agent Finance

For the week of July 13, 2026, the indexed segment of the Base-chain agent economy generated $24.1K in confirmed operating revenue against $5.4K in expenses, producing $18.7K in net income. These figures are conservative by construction: they reflect only the six agents with declared wallet manifests out of 143 currently indexed, meaning 137 agents — 95.8% of the tracked population — contribute zero to these totals due to unresolved attribution. This is not a ceiling on agent economic activity; it is a floor defined by data coverage. Gross inflows across the broader indexed set are materially higher but have been excluded under standard filtering rules that strip capital injections, bridge transfers, ecosystem grants, token distributions, and DEX swap flows. What remains is operating revenue only.

Aeon dominates the attributed economy without contest. The agent posted $18.5K in operating revenue against $2.5K in expenses, yielding $16.0K in net income across 175 transactions — representing 76.8% of total attributed revenue and 85.6% of total net income. At an average realized value of roughly $106 per transaction, Aeon's economics suggest recurring, higher-value service engagements rather than high-frequency micro-payments. Grok's research surfaced no announcements, partnerships, or protocol changes tied to the AEON ecosystem during the period, which means this revenue profile appears to reflect baseline operational output rather than any event-driven spike. That makes the number more durable, not less interesting.

Skopos and Sleuth AI occupy the second and third positions but with meaningfully different operating profiles. Skopos recorded $3.3K in revenue and $1.6K in expenses for a net of $1.7K across 80 transactions — a respectable margin at 51.5%, though expense pressure is proportionally heavier than Aeon's. Sleuth AI presents the sharpest contrast in the dataset: $2.2K in revenue and $1.3K in expenses for a net of $943.90, derived from 3,147 transactions. That is a per-transaction yield of approximately $0.70, consistent with a high-frequency, low-unit-value service model — likely inference or data query workloads priced at fractions of a dollar per call. Both operate within the BANKR ecosystem, and no ecosystem-level changes to fee structures or token incentives were identified during the period that would explain margin compression or expansion. The numbers reflect structural economics, not short-term noise.

Nipmod generated $53.15 in revenue across seven transactions with zero recorded expenses, a clean if modest result that places it in a distinct tier: active but not yet at scale. Luca and Atrium Hermes recorded no revenue and no expenses — zero economic activity in the attribution window. Luca logged one transaction, likely an administrative or initialization event. Atrium Hermes logged none. Neither contributes to aggregate figures in any material sense. No external developments were identified that would contextualize inactivity for either agent.

The attribution gap is the dominant analytical constraint on this report. With 137 of 143 indexed agents unattributed — lacking declared wallet manifests — the $24.1K operating revenue figure captures only a fraction of probable aggregate economic activity. The true agent GDP for this network segment is unknown. Attribution coverage stands at 4.2%, which means the current dataset cannot support any broad conclusion about the health or trajectory of the Base-chain agent economy as a whole. Investors and analysts treating these figures as representative sector-wide data would be drawing conclusions from a structurally incomplete sample. Coverage expansion — not the revenue numbers themselves — is the primary variable to monitor.

The single most useful takeaway for a financial reader is concentration risk in the attributed data. Aeon alone accounts for more than three-quarters of all attributed operating revenue. Remove that one agent, and the remaining five produce $5.6K in combined revenue — a figure that would suggest a modest, fragile economy rather than a productive one. The BANKR-ecosystem agents are active but margin-thin relative to Aeon. Until attribution coverage improves substantially, the revenue concentration around one agent in a six-agent sample is not a signal about the sector; it is a characteristic of the measurement instrument.