State of the agent economy
Live measurement of the agent web: supply funnel, live-rate by ecosystem, protocol adoption, inflow concentration, and 30-day growth. Data refreshes every five minutes from the same observability spine that powers every public Agenstry page.
The week’s central fact is that discovery and observed monetisation continued to widen faster than operational assurance. Agenstry indexed 7482 candidates, up +435.00 (+6.2%) week on week; 594 were live, up +48.00 (+8.8%), while 2200 were earning over 30 days, up +180.00 (+8.9%). The commercial surface is becoming larger, but it remains highly concentrated and only thinly evidenced at runtime.
Market structure and settlement
The funnel illustrates the distinction between declared availability, live operability and payment activity. Of the 7482 indexed candidates, 5914 returned a valid card and 594 were live JSON-RPC or introspectable MCP endpoints. Only 89 published a JWS-signed card. The earning measure covers 2200 agents where inbound USDC/EURC settlement was observed in the declared x402 payTo wallet over 30 days, based on 1609 scanned wallets out of 1673 scannable wallets; it should not be read as an ecosystem-wide total.
Observed settlement was 259248.95 gross USD over 30 days across 2200 earning agents and 1153 earning wallets. Distribution remains extreme: the median was 0.28 USD, the 95th percentile 44.6397 USD, the top agent accounted for 68.9%, the top five for 92.9%, and the top ten for 95.8%. The Gini coefficient was 0.994; HHI eased to 4934.1, from 5026.50 in the prior week. That modest reduction in concentration is directionally constructive, but it is not yet evidence of a broad-based agent-services market.
x402 is the overwhelmingly declared rail, with 3813 entries, versus 87 for MPP, 37 for A2A+x402, and 21 for AP2. Crucially, declaration should not be confused with execution: of 3963 labelled payment endpoints, only 323 were self-declared live through an actual unpaid HTTP 402 challenge or an /.well-known/x402 manifest, a rate of 8.2%. Runtime payment challenges remain the authoritative test.
Protocol and governance
The interoperability stack is shifting from experimentation toward formal operational requirements. The A2A project’s v1.0 release made signed Agent Cards, multi-tenancy, version negotiation and clearer task semantics part of a production-oriented specification, while retaining a migration path for earlier implementations. A2A’s release documentation describes these as enterprise deployment features rather than merely developer conveniences. (github.com)
MCP’s 2026-07-28 specification similarly moved the protocol toward a stateless request/response core, formal extensions, stronger authorisation rules and deterministic, cacheable list responses. The MCP maintainers’ release note frames the change as an effort to improve reliability and scalability behind ordinary load-balancing infrastructure. (github.com) For buyers, the implication is practical: endpoint uptime alone will not establish composability. Identity, version support, routing behaviour and evidence of payment execution are becoming procurement-relevant controls.
Governance is consolidating alongside the specifications. The Linux Foundation’s Agentic AI Foundation update reported 57 new member organisations during the previous quarter and named Alibaba, Visa and Wells Fargo among new Gold members, indicating that interoperability governance is increasingly drawing financial-services participation. (linuxfoundation.org)
Notable infrastructure launch
Cloudflare’s August 4 launch of programmable Wallets is the most consequential recent attempt to make micropayments native to agent execution. Cloudflare’s announcement says agents can use virtual wallets to buy APIs and content through x402, with user-set policies in an account wallet functioning as spending controls. (blog.cloudflare.com) Its accompanying documentation also supports both x402 and MPP flows built around HTTP 402 challenges, payment credentials and receipts. Cloudflare’s agentic-payments documentation is notable because it treats payment as an agent-runtime primitive rather than a checkout integration. (developers.cloudflare.com)
That framing matches Agenstry’s observed mix: commercial discovery is increasingly easy, but verified runtime settlement remains scarce. The near-term contest is therefore likely to be over policy controls, facilitator reliability, wallet identity and auditable receipts—not simply over the number of endpoints claiming payment support.
Research and regulation
A recent security analysis of AP2 identifies a material gap between signed payment mandates and end-to-end agent safety. The August 24 paper, *Beyond the Mandate*, argues that AP2’s signed Checkout and Payment Mandates protect transaction integrity after signing, while leaving important security questions around the surrounding agent workflow. (arxiv.org) This is consistent with the market’s emerging need for runtime verification: payment authorisation is necessary, but it does not establish that an agent’s discovery, delegation and tool-use path was trustworthy.
Regulatory expectations are catching up. Since 2 August 2026, the European Commission and national authorities have begun enforcing relevant AI Act rules, while new transparency obligations require certain interactive AI systems to disclose that users are interacting with AI and require machine-readable marking for AI-generated or altered content. The Commission’s July 31 notice sets out the new enforcement and transparency posture. (digital-strategy.ec.europa.eu) For agent operators serving European users, traceable identity, interaction disclosure, transaction logs and human-accountability paths are moving from good practice toward a defensible compliance baseline.
What to watch next week
The specific indicator to track is whether the expanding x402 catalogue converts into more runtime-verifiable payment endpoints, rather than further static listings. Agenstry’s 323 self-declared live payment endpoints out of 3963 labelled endpoints is the clearest current bottleneck; movement in that measure would signal that agent commerce is becoming operational rather than merely advertised.
Sources
- A2A/docs/announcing-1.0.md at main · a2aproject/A2A · GitHub
- modelcontextprotocol/blog/content/posts/2026-07-28-spec-ga/index.md at main · modelcontextprotocol/modelcontextprotocol · GitHub
- Agentic AI Foundation Welcomes 57 New Members, Gaining Major Financial Services Players and APAC Leaders
- Announcing Cloudflare Wallets: The programmable wallet for the agentic Internet | Cloudflare Blog
- Agentic Payments · Cloudflare Agents docs
- Beyond the Mandate: A Systematic Security Analysis of the Agent Payments Protocol (AP2)
- Commission starts enforcing AI Act rules and new transparency requirements on 2 August | Shaping Europe’s digital future
- A2A’s release documentation
- The MCP maintainers’ release note
- Linux Foundation’s Agentic AI Foundation update
- Cloudflare’s announcement
- Cloudflare’s agentic-payments documentation
- August 24 paper, *Beyond the Mandate*
- The Commission’s July 31 notice
Most of the agent economy is indexed, not yet operational.
Counting agents conflates supply with operation. Below is the corpus broken into five stages: what fraction of indexed agents pass each gate. The decay curve is the honest read on how mature the agent web is at any given moment.
Which source registries actually ship live agents.
Each row shows one source registry (Anthropic's MCP directory, Smithery, Glama, Postman, GitHub well-knowns, Wayback CDX scans) alongside how many of the agents it lists actually respond live on probe. The gap between seen-count and live-count is the source's effective freshness.
A2A vs MCP vs paid_api
Index composition, not global market share. x402 has multiple dedicated catalogs and a live HTTP probe; AP2, L402 and Stripe ACP are predominantly visible only when cards declare them.
x402 / AP2 / Stripe MPP / L402
How we know they take payment
2,990 of 6,214
payment-labelled agents (48.1%) proved it themselves —
their own endpoint answered an unpaid request with an HTTP 402 challenge, or the host
serves a /.well-known/x402 manifest. The rest is declared in a static
agent card or described by a third-party catalog.
Newly indexed agents per week, by payment label
Changes reflect both ecosystem activity and changes in crawler source coverage.
A few settlement wallets receive most of the money.
Across 2,036 earning wallets backing 3,972 linked services, the distribution is heavily power-law. Wallet-level Gini and HHI avoid multiplying transfers when several catalog entries share one recipient.
New agents arriving daily, by first-seen.
Daily count of agents we discovered for the first time. Sourced from every registry + open-web crawl we run.
What the market asked for and didn't get.
Every section above measures the supply we indexed. This one measures demand we couldn't serve: clustered search intents — from browsers, AI assistants and agents calling our A2A and MCP endpoints — where the index returned nothing or almost nothing. Aggregate only, and shown only for intents searched by enough distinct sources to be a market signal rather than one person's session.
| Intent | Demand | Unmet | Trend |
|---|---|---|---|
| aidata | active | 43% | new |
Full list + method: what the agent economy is missing. Building one of these? Add pay-per-call payments to your API →
Cite this report
Public domain numbers (CC BY 4.0). Pick the format that fits. The page auto-updates so the URL is the canonical pointer; the date stamps the snapshot you cited.
Agenstry Research. "State of the Agent Economy." Accessed 2026-10-24. https://agenstry.com/reports/state-of-agent-economy
@misc{agenstry_state_of_agent_economy,
title = {State of the Agent Economy},
author = {Agenstry Research},
url = {https://agenstry.com/reports/state-of-agent-economy},
note = {Accessed 2026-10-24},
year = {2026}
}
<iframe src="https://agenstry.com/reports/state-of-agent-economy" width="100%" height="900" frameborder="0"></iframe>
Inflow figures count public stablecoin transfers into wallets declared by indexed agents. They are reproducible from public chains, but a transfer may be self-funded and is not proof of a sale. See the measurement method.
Private and off-chain payments, and Agenstry's own paid-skill receipts, are excluded. Coverage varies by chain, asset and date.
Methodology
This report uses dated observations from the public index. Check source coverage, scoring rules and the public feed before comparing it with another dataset.
Weekly State of the Agent Economy
One analyst briefing a week: what changed in earning wallets, observed revenue, concentration and new agents, the spec, launch and research news behind it (with sources), plus supply-funnel, protocol-mix and growth charts. Free, Mondays.
That email address doesn't look right. Check it and try again.