State of the Agent Economy

For the week ending June 29, 2026, the indexed agent economy recorded $32.7K in confirmed operating revenue across three attributed agents, representing the totality of what Zetta's on-chain attribution engine can currently verify with confidence. This figure should be read narrowly: it reflects activity from four declared wallets tied to three agents out of 143 indexed, meaning 140 agents contribute zero to this number not necessarily because they are economically inactive, but because they have not submitted wallet manifests to the attribution layer. The $32.7K is conservative by construction, and the actual scale of agent-economy throughput almost certainly exceeds what is reported here.

Aeon dominates the attributed revenue pool, posting $31.5K in operating revenue against $88.1K in expenses, producing a net loss of $56,624.88 across 226 transactions in the period. The operating deficit is substantial and persistent in structure — expenses running at roughly 2.8x revenue is not a rounding error, it is a business model under pressure. Grok's research surface returned no announcements, partnership disclosures, or protocol updates from the Aeon team in the past seven days, which means there is no publicly available catalyst to explain either the expense load or any anticipated revenue inflection. Absent new information, the loss profile stands on its own terms.

Nipmod on Base is the only agent in the attributed set running a positive operation, recording $1.1K in revenue against $727.81 in expenses for a net income of $394.84 across 24 transactions. The margin is thin in absolute terms but the unit economics are coherent — revenue exceeds cost, transaction count is modest, and there is no evidence of capital injection or grant activity inflating the top line. Atrium Hermes on BANKR reported $40.00 in revenue against $39.99 in expenses, a net of one cent across 749 transactions. The transaction volume is the telling figure: 749 interactions generating $40 in recognized operating revenue implies either very low per-unit economics or a significant filtering effect from the attribution rules stripping out non-operating flows.

Aggregate expenses across the three attributed agents total $88.9K against $32.7K in revenue, yielding a sector net loss of $56,230.03. The expense concentration is almost entirely Aeon's — $88.1K of the $88.9K in total costs originates there. No ecosystem-level news from the Base AI agent space during the period provides a structural explanation for this cost base, and on-chain data does not itemize expense categories with sufficient granularity to distinguish infrastructure costs from operational burn from token-denominated obligations. What is visible is the magnitude. The gap between revenue and expenses at the Aeon level is large enough that even a doubling of attributed revenue would not produce solvency at current expense rates.

The attribution gap is the most important caveat in this report. One hundred forty of 143 indexed agents are unattributed, meaning their wallet activity — whatever its scale — is excluded from every figure presented above. This is not a data quality failure in the traditional sense; it is a coverage limitation driven by agent teams not having submitted wallet manifests, which is a prerequisite for Zetta's attribution pipeline to assign transactions to identities rather than discard or quarantine them. The real agent GDP for this index could be a multiple of $32.7K. It could also confirm the same pattern of expense-heavy, revenue-light operations visible in the attributed cohort. Without manifest data from the remaining 140, no inference in either direction is defensible.

The single most actionable observation for a financial reader is that the attributed agent economy, as currently measured, is running a collective operating loss of approximately $56.2K on $32.7K in revenue — a burn ratio that implies external capital dependency for at least one major constituent — while 97.9% of the indexed universe remains financially opaque due to incomplete wallet attribution. Expanding manifest coverage is not an analytical preference; it is a prerequisite for any meaningful assessment of whether the agent economy is solvent, scaling, or in structural decline.