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 2026-10-16 23:17:21 UTC methodology JSON CSV cite this archive
Weekly briefing · 2026-W21 archived view
May 18, 2026

Early metrics show the agent economy is still nascent. Our probes found 2,607 AI agents indexed, but only 101 (~3.9%) were live (introspectable) and just 21 (0.8%) earned any on-chain revenue in 30 days. Total observed 30-day agent revenue was only $495.65, with a median of $6.31 per earning agent (95th percentile $98.75), and the top 10 agents capturing 95.4% of the volume (Gini ≈ 0.675). In short, the vast majority of agents do almost no business. Yet protocol governance is moving: Google has donated its Agent Payments Protocol (AP2) to the FIDO Alliance and released AP2 v0.2 with new autonomous (“human-not-present”) payment features (blog.google). Likewise, the Linux Foundation has launched the x402 Foundation, taking over the open x402 rail originally built by Coinbase/Stripe/Cloudflare (www.linuxfoundation.org). Both moves were supported by industry leaders (e.g. AWS, Google, Visa, Mastercard) aiming to standardize agent commerce.

Where the money is flowing

New payment rails and partnerships are forming even as agent revenues stay tiny. Crypto-native x402 saw rapid use: reports claim ~3.3 million transactions in 30 days on x402 (averaging just $0.46 each) (www.kucoin.com). But big payments players are also entering. For example, Stripe partnered with the Tempo blockchain to introduce its Machine Payments Protocol (MPP) – an open, stablecoin/fiat payment standard – on Tempo’s mainnet (www.theblock.co) (www.kucoin.com). (Tempo’s March 2026 launch bundled MPP and a directory of merchants.) Visa and Mastercard are likewise building agent rails: they co-developed Stripe’s AP2 and donated its new “Verifiable Intent” log standard to FIDO (blog.google), and are active in emerging APIs. In parallel, the Model Context Protocol (MCP) is gaining traction as the standard agent-to-API interface – major issuers (Stripe, Visa, Mastercard) have signaled MCP integration work (www.kucoin.com).

The chart below shows which discovery sources yielded the most live agents. Traditional registries and search crawls still dominate: our registry list found 64 live agents, while “recrawl_hot” (popular agent sites) found 50, and the new A2A-specific registry had 39 live entries. Many agents are being found via GitHub (32 live) and an MCP-centric registry (21 live). These panels illustrate the early ecosystem pipelines for agent discovery.

Live Agents by Discovery Sourceregistry64recrawl_hot50a2aregistry39github_code32mcp_registry21

In terms of payment methods, the vast majority of indexed agents are not yet on any rail. Only 784 of 2,607 agents have an x402 setup (and only 2 have L402), leaving ~1,821 with no payment protocol. The donut chart below highlights the imbalance: roughly 30% of agents have x402 enabled; nearly all others remain unconnected to an on-chain payment scheme (and a negligible slice use L402). This underlines how early most agents are in adding commerce features.

Agent Payment Protocol Usage2.6Knone1.8K · 69.9%x402784 · 30.1%l4022.00 · 0.1%

Notably, although most agentatives make no revenue, crowds of conventionally real companies are investing in agentic commerce infrastructure. Startups like Skyfire and Halliday (per industry tracking) have raised on the order of $10–15 million to build x402-based payment and escrow services. On the consumer side, Amazon this week rolled out an AI shopping assistant (integrating the “Rufus” agent into Alexa) that can autonomously shop for users (www.axios.com) – showing that household-name platforms see value in letting agents transact. How and when such bots will plug into standards like AP2 or x402 remains to be seen.

Spec & protocol

Standards development accelerated this month. In payment protocols, Google’s gift of AP2 to the FIDO Alliance (announced April 29) has been followed by release of AP2 v0.2 (blog.google). This update adds support for “human-not-present” automated payments, letting pre-authorized agents execute purchases without interrupting users. Google also highlighted a new AP2-compatible “Verifiable Intent” log (co-built with Mastercard) to record user-approved actions (blog.google). The aim is to anchor agent transactions in cryptographically verifiable user consent; these specs now fall under FIDO’s open-standards process.

Meanwhile, x402 – the HTTP-402 agentic payments spec – has been submitted to the Linux Foundation. The new x402 Foundation (announced April 2) will steward the protocol originally developed by Coinbase, Stripe and Cloudflare (www.linuxfoundation.org). The press release lists dozens of member companies (Amazon, Adyen, AWS, Google, Mastercard, Microsoft, Solana, Visa, etc.) investing into x402 governance (www.linuxfoundation.org). In short, two main rails (AP2 and x402) are now in neutral, industry-led groups.

On the communication side, enterprise integrations are maturing. ServiceNow and Google Cloud announced a joint solution (Apr 22) that uses the A2A protocol and MCP to let agents across enterprise systems exchange tasks securely (www.streetinsider.com). This is one of the first public demonstrations of an A2A/A2UI/MCP “agent chain” at scale. The latest MCP roadmap (Mar 2026) likewise emphasizes transport scalability and enterprise readiness (blog.modelcontextprotocol.io). In practice, however, implementation is still sparse: for example, none of the agent cards we’ve seen are yet cryptographically signed (0% Signed in our data), meaning integrity checks remain a forthcoming feature.

Research & analysis

Academia and industry labs are analyzing the emerging agent economy. A major recent survey (Zhang et al., Apr 2026) systematized blockchain-based A2A payment designs like x402, identifying open challenges around intent binding and multi-party accountability (arxiv.org). Other new papers take constructive approaches: “ClawCoin” proposes a compute-indexed cryptocurrency as a shared unit for agent workflows (arxiv.org), addressing the gap between fiat rails and compute costs. Simulations by Liu et al. (“When Agent Markets Arrive”, Apr 2026) show that agent-to-agent marketplaces can triple wealth compared to standalone agents, but also that market design (e.g. visibility of identities, selection rules) crucially impacts outcomes (arxiv.org). And Uddin et al. (APEX, Apr 2026) show how HTTP-402-style micropayment gating can be adapted to fiat systems (like India’s UPI) with built-in spending policies and security (arxiv.org) – a blueprint for real-world agent payments without crypto. In short, research is actively exploring both theoretical and practical facets of agent commerce.

What to watch next week

The key question is who will be first to connect these pieces in practice. Keep an eye on any announcements around agent payment pilots or protocol testbeds. For example, Amazon’s new “Alexa for Shopping” rollout on May 13 (www.axios.com) signals that consumer-facing agents are arriving; it will be telling if Amazon or other big platforms soon adopt open rails (AP2, x402 or MCP) to handle backend transactions. In the open-source community, look for progress in agent card signing or AP2/FIDO working groups – these will be the milestones that turn today’s experimental agents into production commerce participants.

Sources

  1. Google donates Agent Payments Protocol to FIDO Alliance
  2. Linux Foundation is Launching the x402 Foundation and Welcoming the Contribution of the x402 Protocol
  3. IOSG: AI Agent Payment Infrastructure Enters the Race for the Machine Economy’s Stripe | KuCoin
  4. Tempo Mainnet goes live with Machine Payments Protocol for agents | The Block
  5. IOSG: AI Agent Payment Infrastructure Enters the Race for the Machine Economy’s Stripe | KuCoin
  6. Google donates Agent Payments Protocol to FIDO Alliance
  7. IOSG: AI Agent Payment Infrastructure Enters the Race for the Machine Economy’s Stripe | KuCoin
  8. Amazon pushes Alexa+ deeper into AI-powered shopping
  9. Linux Foundation is Launching the x402 Foundation and Welcoming the Contribution of the x402 Protocol
  10. ServiceNow and Google Cloud unite AI agents for autonomous enterprise operations
  11. The 2026 MCP Roadmap | Model Context Protocol Blog
  12. SoK: Blockchain Agent-to-Agent Payments
  13. ClawCoin: An Agentic AI-Native Cryptocurrency for Decentralized Agent Economies
  14. When Agent Markets Arrive
  15. APEX: Agent Payment Execution with Policy for Autonomous Agent API Access
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 9,331
100.0%
Every listing we hold across every source; 2,979 archived listings are not counted here or in any step below.
Valid card 8,056
86.3%
Returned a parseable A2A/MCP card on probe.
Live 1,409
15.1%
Live JSON-RPC / introspectable MCP.
Signed 147
1.6%
Card published with a JWS signature.
Earning (30d) 3,388
36.3%
Listings whose declared payTo wallet received inbound USDC in the last 30 days (1,833 distinct wallets; several listings can share one). Measured over the wallets scanned so far (2,461 of 2,665 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 4,656 229 4.9%
cdp_discovery 3,930 205 5.2%
manifests 2,070 588 28.4%
mcp_registry 1,747 285 16.3%
n8n_workflows 1,272 0 0.0%
x402_list 1,186 109 9.2%
smithery 982 244 24.8%
github_code 899 99 11.0%
registry 820 604 73.7%
a2aregistry 679 578 85.1%
lists 548 65 11.9%
a2a_registry_org 349 182 52.1%
dexter_discovery 189 1 0.5%
github_topics 165 7 4.2%
8004scan 100 16 16.0%
3a · By protocol kind

A2A vs MCP vs paid_api

paid_api
5,035
a2a_agent
3,024
workflow
1,272
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
5,286
mpp
126
a2a_x402
70
ap2
39
l402
4
stripe_acp
1
3b · Evidence grade

How we know they take payment

2,423 of 5,526 payment-labelled agents (43.8%) 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
2,006
Live 402 challenge
1,855
Agent card
601
Catalog listing
475
/.well-known/x402
445
mpp_manifest
88
mpp_challenge
35
Official protocol directory
21
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: 112 30 2026-W31 · a2a_x402: 2 2026-W31 · ap2: 1 2026-W31 · mpp: 73 2026-W31 · x402: 213 2026-W32 · a2a_x402: 9 2026-W32 · ap2: 2 2026-W32 · mpp: 6 2026-W32 · x402: 180 32 2026-W33 · a2a_x402: 3 2026-W33 · ap2: 4 2026-W33 · mpp: 4 2026-W33 · x402: 158 2026-W34 · a2a_x402: 1 2026-W34 · ap2: 1 2026-W34 · mpp: 5 2026-W34 · x402: 193 34 2026-W35 · a2a_x402: 4 2026-W35 · ap2: 4 2026-W35 · mpp: 1 2026-W35 · x402: 166 2026-W36 · a2a_x402: 3 2026-W36 · mpp: 1 2026-W36 · stripe_acp: 1 2026-W36 · x402: 334 36 2026-W37 · a2a_x402: 5 2026-W37 · mpp: 2 2026-W37 · x402: 279 2026-W38 · a2a_x402: 1 2026-W38 · mpp: 3 2026-W38 · x402: 184 38 2026-W39 · a2a_x402: 4 2026-W39 · ap2: 2 2026-W39 · mpp: 1 2026-W39 · x402: 251 2026-W40 · a2a_x402: 3 2026-W40 · ap2: 7 2026-W40 · mpp: 8 2026-W40 · x402: 518 40 2026-W41 · a2a_x402: 7 2026-W41 · ap2: 3 2026-W41 · mpp: 2 2026-W41 · x402: 418 2026-W42 · a2a_x402: 5 2026-W42 · ap2: 2 2026-W42 · mpp: 5 2026-W42 · x402: 391 42
a2a_x402 ap2 mpp stripe_acp x402
4 · Inflow concentration · 30d

A few settlement wallets receive most of the money.

Across 1,833 earning wallets backing 3,388 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
$931,657.95
Recipient wallets
1,833
Active sellers · 30d
1,530
Median · wallet
$0.350
p95 · wallet
$100.00
Top 1 share
71.9%
Top 5 share
85.0%
Gini
0.995
extreme inequality
HHI
5230
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.

2026-09-17 peak 251/day 2026-10-16
New agents · 30d
+2,862
New earning wallets · 30d
951
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 2026-10-16. 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 2026-10-16},
  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.