Skip to content
Report · auto-updating

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.

As of 2026-10-11 22:10:03 UTC methodology JSON CSV cite this archive
Weekly briefing · 2026-W29 archived view
Jul 13, 2026

Agenstry’s weekly data show another uptick in discovered agents but virtually no change in actual commerce. We indexed 4,491 agent endpoints (a 6.0% jump from 4,235 a week earlier), of which 276 (6.1%) were live MCP servers【data】. Yet only 21 agents (0.5% of indexed) saw on-chain revenue in the last 30 days (unchanged WoW)【data】. These 21 agents generated $491.02 total (median $6.31), with the top five earning 79.2% of it【data】 and the top ten 95.1% (Gini 0.683, HHI 1495.8). In infrastructure, major players doubled down on enabling agent payments: Stripe’s new Machine Payments Protocol (MPP) – co-developed with fintech Tempo – provides an “internet-native” payment standard for AI agents (stripe.com), and Cloudflare in early July launched an x402-based “Monetization Gateway” to charge for any API or data asset, settling all fees in stablecoins (blog.cloudflare.com). Together these moves (along with the newly announced x402 Foundation consortium) signal growing support for stablecoin rails in the agent economy.

Where the money is flowing

Despite rising discovery, the agent economy’s cash flows remain minuscule and extremely concentrated. In the past 30 days only 21 agents earned at all, splitting $491.02 total – a trivial volume where the median take is just $6.31【data】. The distribution is highly top-heavy (top1 share 21.7%, top5 79.2%, top10 95.1%【data】). These funds flow primarily in stablecoins: for example, Stripe’s Machine Payments Protocol explicitly supports on-chain payments (via stablecoins or fiat) for agent services (stripe.com) (stripe.com), and Cloudflare’s new gateway settles authorizations in USD-pegged tokens (blog.cloudflare.com). Anecdotally, companies are already building on these rails – Stripe notes that services like “Browserbase” let agents spin up browsers and pay per session via MPP (stripe.com). Absent big consumer launches, most investment has been in infrastructure: staffing, wallets, and agent platforms. (For comparison, non-agentic giants continue ramping AI pay capabilities – e.g. Amazon Bedrock’s preview of agentic payments with Coinbase/Stripe in May – but no new fundraising or IPOs tied specifically to agents were reported this week.)

Daily new agents added to index05141.0K2026-06-142026-07-13

Chart: Daily new agents registered in 30 days (note the single-day spike on July 6).

Spec & protocol

Major standards continue evolving. The Model Context Protocol (MCP) maintainers have published a “2026-07-28” release candidate, introducing a stateless core (breaking existing session-based assumptions), a formal extension framework (Tasks → extension, plus new “MCP Apps”), hardened auth checks and a deprecation policy (blog.modelcontextprotocol.io). This is the largest MCP rewrite to date, and its final spec is on track for July 28 (after a ten-week SDK validation period) (blog.modelcontextprotocol.io). In payments, the Agent-to-Agent standard (x402) is nearing formal adoption: for example, an IETF Internet-Draft co-authored by Stripe/Tempo (“Payment HTTP Authentication Scheme”) was published in March to regularize the agentic 402 challenge/response flow (datatracker.ietf.org). Our agent index shows x402 dominating (525 of 529 cards), with only 3 using the older “l402” fork and 1 using Google’s AP2【data】. Cloudflare has also embedded x402 support directly into its Agents SDK and Workers ecosystem, providing an “x402-proxy” template that automatically returns a 402 challenge on protected endpoints (developers.cloudflare.com). (By contrast, adoption of agent-card signing remains rare: only 31 cards are JWS-signed, ~0.7% of indexed agents【data】.)

Agent payment protocols529x402525 · 99.2%l4023.00 · 0.6%ap21.00 · 0.2%

Chart: Payment protocols in agent cards (nearly all use x402, as captured in our index【data】).

Notable launches

Cloudflare dominated this week’s announcements. On July 1 it launched a Monetization Gateway – a global edge service that lets any Cloudflare-protected URL (web page, API, even an MCP tool endpoint) require an x402 payment for access (blog.cloudflare.com). All transactions initially settle in USDC (stablecoin) over x402 (blog.cloudflare.com). The Gateway includes a flexible rules API and global Worker templates (e.g. an “x402-proxy” to attach payment checks in front of any endpoint) so that charging agents becomes a configuration task rather than custom dev. Cloudflare says its waitlist is open and aims for sub-second settlement worldwide (blog.cloudflare.com) (blog.cloudflare.com). (Outside of agent-specific projects, broader AI infrastructure news included AWS’s May preview of Bedrock AgentCore payments with Coinbase/Stripe, but no major new launches were reported this week.)

Agent pipeline drop-offIndexed4.5KValid3.0KLive276Signed31Earning21

Chart: Drop-off from indexed candidates to earning agents. Only 0.5% of spotted agents show any revenue【data】. (The “Signed” row is agent cards carrying a JWS signature.)

Research

Academic interest in multi-agent AI remains brisk. A new arXiv study found that teams of LLM agents “substantially outperform human teams in creativity” on brainstorming tasks (Cohen’s d≈1.50, comparing 4,541 AI-generated ideas vs 341 human-generated ones) (voiceclone5090.helport.ai). Another group introduced Emergence World, a continuous simulation platform that runs hundreds of LLM-driven agents in a shared virtual world (fed with real weather, news APIs, etc.), to measure long-term dynamics like behavioral drift and governance (huggingface.co). More foundational surveys are appearing too; e.g. one “Hitchhiker’s Guide to Agentic AI” has been posted, aiming to cover the full stack of autonomous agents from theory to deployment. Such work underscores both the promise and complexity of scaling AI agents, but shows that in controlled tests multiple agents can synergize to solve problems beyond a single model’s capability (voiceclone5090.helport.ai).

Regulation

Policy moves in AI and crypto are starting to touch agents. In the EU, on June 29 the Council finalized a new AI regulation that “streamlines and simplifies” existing rules under a single framework (www.consilium.europa.eu) (an “omnibus” package to cut red tape; the impact on agentic systems remains broad and non-specific). Financial regulators in the US have been more pointed: in mid-June FinCEN/OCC proposed treating fiat-backed stablecoin issuers like banks (per the GENIUS Act), requiring them to run full KYC/Customer Identification Programs (www.fincen.gov) (cryptoslate.com). Crypto analysts note this effectively installs identity checks “at the issuer gate” for fiat-to-stablecoin on-ramps, while not extending those requirements into the peer-to-peer transfer layer (cryptoslate.com). No agent-specific law was passed this week, but these developments set the rules of the game for how agents buy assets.

What to watch next week

With the MCP stateless spec set to finalize on July 28, the community will be watching whether major SDKs ship updates in time and whether any interoperability issues arise. The market will also track early usage of Cloudflare’s gateway: as of W29 it’s in private beta/waitlist, so any spikes in x402 transaction volume would be a leading indicator. On the governance side, we expect continued clarity on how regulators apply AML rules to automated agents (e.g. stablecoin KYC); comments on the proposed rules close in late summer. In research, upcoming AI conferences (like NeurIPS workshops) may unveil further multi-agent benchmarks or systems, which could influence the next cycle of the agent ecosystem.

Sources

  1. Introducing the Machine Payments Protocol
  2. Announcing the Monetization Gateway: charge for any resource behind Cloudflare via x402
  3. Introducing the Machine Payments Protocol
  4. The 2026-07-28 MCP Specification Release Candidate | Model Context Protocol Blog
  5. draft-ryan-httpauth-payment-01 - The "Payment" HTTP Authentication Scheme
  6. Charge for HTTP content · Cloudflare Agents docs
  7. Announcing the Monetization Gateway: charge for any resource behind Cloudflare via x402
  8. Announcing the Monetization Gateway: charge for any resource behind Cloudflare via x402
  9. Multi-agent AI systems outperform human teams in creativity · AI-Map
  10. Paper page - Emergence World: A Platform for Evaluating Long-Horizon Multi-Agent Autonomy
  11. Inteligência artificial: Conselho dá luz verde final à simplificação e racionalização das regras - Consilium
  12. FinCEN, Agencies Propose Rule to Implement GENIUS Act Customer Identification Program Requirement | FinCEN.gov
  13. US starts clock to bring in ID checks for converting dollars to stablecoins but DeFi stays outside the rules
  14. US starts clock to bring in ID checks for converting dollars to stablecoins but DeFi stays outside the rules
1 · Supply funnel

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.

Stage Count % of indexed Definition
Indexed 8,901
100.0%
Every listing we hold across every source; 2,852 archived listings are not counted here or in any step below.
Valid card 7,673
86.2%
Returned a parseable A2A/MCP card on probe.
Live 1,219
13.7%
Live JSON-RPC / introspectable MCP.
Signed 144
1.6%
Card published with a JWS signature.
Earning (30d) 3,206
36.0%
Listings whose declared payTo wallet received inbound USDC in the last 30 days (1,692 distinct wallets; several listings can share one). Measured over the wallets scanned so far (2,290 of 2,549 with a wallet) — not an ecosystem-wide total. Not a subset of Live: most are paid APIs, which answer 402 rather than a free JSON-RPC probe.
2 · Live-rate by ecosystem

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.

Source Seen Live Live rate
agentic_market 4,324 193 4.5%
cdp_discovery 3,634 170 4.7%
manifests 2,068 540 26.1%
mcp_registry 1,747 280 16.0%
n8n_workflows 1,226 0 0.0%
x402_list 1,088 85 7.8%
github_code 894 88 9.8%
smithery 855 197 23.0%
registry 775 481 62.1%
a2aregistry 641 457 71.3%
lists 547 58 10.6%
a2a_registry_org 320 152 47.5%
dexter_discovery 182 1 0.5%
github_topics 163 8 4.9%
mpp 93 1 1.1%
3a · By protocol kind

A2A vs MCP vs paid_api

paid_api
4,760
a2a_agent
2,916
workflow
1,226
3b · By observed payment-protocol label

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

x402
4,981
mpp
124
a2a_x402
64
ap2
36
l402
4
stripe_acp
1
3b · Evidence grade

How we know they take payment

2,144 of 5,210 payment-labelled agents (41.2%) 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.

Facilitator feed
1,977
Live 402 challenge
1,629
Agent card
578
Catalog listing
483
/.well-known/x402
402
mpp_manifest
81
mpp_challenge
32
Official protocol directory
28
3c · Indexed protocol labels over time · last 12 weeks

Newly indexed agents per week, by payment label

Changes reflect both ecosystem activity and changes in crawler source coverage.

2026-W30 · a2a_x402: 1 2026-W30 · x402: 243 30 2026-W31 · a2a_x402: 2 2026-W31 · ap2: 1 2026-W31 · mpp: 81 2026-W31 · x402: 215 2026-W32 · a2a_x402: 9 2026-W32 · ap2: 2 2026-W32 · mpp: 2 2026-W32 · x402: 186 32 2026-W33 · a2a_x402: 3 2026-W33 · ap2: 4 2026-W33 · mpp: 4 2026-W33 · x402: 156 2026-W34 · a2a_x402: 1 2026-W34 · ap2: 1 2026-W34 · mpp: 5 2026-W34 · x402: 198 34 2026-W35 · a2a_x402: 4 2026-W35 · ap2: 4 2026-W35 · mpp: 1 2026-W35 · x402: 192 2026-W36 · a2a_x402: 3 2026-W36 · mpp: 1 2026-W36 · stripe_acp: 1 2026-W36 · x402: 338 36 2026-W37 · a2a_x402: 5 2026-W37 · mpp: 3 2026-W37 · x402: 283 2026-W38 · a2a_x402: 1 2026-W38 · mpp: 3 2026-W38 · x402: 190 38 2026-W39 · a2a_x402: 4 2026-W39 · ap2: 2 2026-W39 · mpp: 1 2026-W39 · x402: 262 2026-W40 · a2a_x402: 3 2026-W40 · ap2: 7 2026-W40 · mpp: 7 2026-W40 · x402: 536 40 2026-W41 · a2a_x402: 7 2026-W41 · ap2: 2 2026-W41 · mpp: 2 2026-W41 · x402: 413 41
a2a_x402 ap2 mpp stripe_acp x402
4 · Inflow concentration · 30d

A few settlement wallets receive most of the money.

Across 1,692 earning wallets backing 3,206 linked services, the distribution is heavily power-law. Wallet-level Gini and HHI avoid multiplying transfers when several catalog entries share one recipient.

Observed inflow · 30d
$959,643.00
Recipient wallets
1,692
Active sellers · 30d
1,434
Median · wallet
$0.380
p95 · wallet
$108.79
Top 1 share
69.8%
Top 5 share
83.2%
Gini
0.994
extreme inequality
HHI
4938
highly concentrated
5 · 30-day growth

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.

2026-09-12 peak 251/day 2026-10-11
New agents · 30d
+2,537
New earning wallets · 30d
795
first observed earning in the last 30 days

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.

Plain text
Agenstry Research. "State of the Agent Economy." Accessed 2026-10-11. https://agenstry.com/reports/state-of-agent-economy
BibTeX
@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 2026-10-11},
  year   = {2026}
}
Embed (iframe)
<iframe src="https://agenstry.com/reports/state-of-agent-economy" width="100%" height="900" frameborder="0"></iframe>
Scope What this report measures, and what it doesn't

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.