How many x402 transactions have been recorded?

Agent Economy reported 186,554,450 cumulative x402 transactions across its tracked networks on 11 September 2026. Coverage and methodology differ between trackers, so the number should be cited with its source and date rather than treated as a universal protocol total.

How much x402 volume has been settled?

Agent Economy reported $41,721,405 of cumulative x402 settled value on 11 September 2026 across its tracked coverage. That is a raw settlement measure, not verified business revenue or a trailing-30-day organic-commerce figure.

Some x402 settlement activity represents genuine economic activity, but raw totals can also include tests, self-directed usage, dust and infrastructure flows. x402stats currently estimates about 86% of the raw trailing-30-day volume in its own filtered dataset as organic, while other historical studies have applied different filters and produced different ratios. Always cite the tracker, date, window and methodology.

Is x402 widely adopted on the public web?

Not yet by top-domain count. Agent Economy's reading of Cloudflare Radar's week-of-31-August agent-readiness scan found x402 on 6 of 107,002 successfully scanned relevant top domains. That scan is not a census of every API or subdomain, and x402 is only applicable to sites exposing paid resources.

Does a high x402 transaction count prove strong commercial demand?

No. Transaction count measures activity, not buyer quality or sustainable demand. Commercial evidence is stronger when independent buyers repeatedly pay for useful services, receive delivery and return without seller-funded activity.

Are most x402 listings getting repeat buyers?

Not in the independent snapshots reviewed here. A 28 July survey of Coinbase Bazaar found 520 of 14,865 listings, or 3.5%, passed the author's repeat-demand filter of at least three unique payers and at least three calls per payer on average. A separate August x402stats report counted 124 seller wallets as organic under a different revenue-and-buyer threshold. These are third-party methods, not protocol-wide ground truth, but both show that listing volume should not be mistaken for repeat demand.

Why can an x402 API be listed but still have no buyers?

A listing solves only part of discovery. A buyer still needs a concrete job, understandable input and output, a price it can justify, a reachable endpoint, compatible payment terms and enough trust to spend. Independent Bazaar research found repeat usage concentrated in established search and data services, suggesting that useful outcomes matter more than simply exposing an x402 paywall.

Editorial

x402 stats · September 2026

x402 stats: 186M+ transactions and $41.7M settled — but raw volume is not buyer demand.

The protocol can show substantial settlement activity while remaining rare on the public web and highly concentrated at the seller layer. This current snapshot separates transactions, settled value, organic-volume estimates, public-web exposure, listings and repeat-buyer evidence so builders do not mistake one signal for another.

Published 9 September 2026 · updated 11 September 2026 · source values current to the dates named below

Cumulative transactions

186.55M

Agent Economy · 11 Sep

Settled value

$41.72M

Agent Economy · 11 Sep

Daily x402 transactions

14,516

Agent Economy · 11 Sep

Top-domain detections

6 / 107K

Radar-derived · week of 31 Aug

Signal 01

Settlement activity is real, but activity is not adoption

Agent Economy reported 186,554,450 cumulative x402 transactions and $41,721,405 of cumulative settled value on 11 September across its tracked coverage. The simple cumulative ratio is roughly 22 cents per transaction. It is not a median service price and does not tell us how much activity came from recurring commercial workflows.

The same dataset reported 14,516 transactions for 11 September. Daily activity can move sharply, so one-day counts are a weak basis for declaring either mass adoption or collapse.

Interpretation

Count the economic loop, not just the payment

Discover → understand → select → pay → deliver → reuse.

A durable market needs independent buyers completing that loop because the service solves a job they actually have.

HTTP 402 responses, test calls, incentive traffic and seller-funded purchases should not be promoted into revenue evidence.

Which number should you cite?

There is no single x402 volume number without a scope and methodology

For a dated cumulative on-chain snapshot, Agent Economy currently reports 186.55 million tracked x402 transactions and $41.72 million settled as of 11 September 2026, with its data refreshed from public chain activity.

For a filtered trailing-30-day view, x402stats currently says roughly 86% of the raw volume in its own dataset is organic — about $1.2 million of $1.3 million — after removing the wash, dust and facilitator patterns defined in its methodology. That is a different question from cumulative protocol settlement and should not be blended into the same headline.

Be cautious with the widely repeated x402.org “Last 30 Days” panel showing 75.41 million transactions and $24.24 million. A 6 September audit documented the same four headline values across repeated checks from 27 August through 6 September and noted that the panel exposed no last-updated timestamp or methodology. That does not establish why the values were unchanged; it does mean they should not be treated as a fresh rolling-window measurement without a newer timestamped source.

Agent Economy live x402 stats →x402stats organic-volume methodology →September audit of the x402.org counter →

Signal 02

The web-adoption gap is still large

Cloudflare Radar now measures AI-agent standards across successfully scanned top web domains. Agent Economy's 5 September extraction of the week-of-31-August scan found x402 on only 6 of 107,002 successfully scanned relevant domains. By comparison, it reported 361 MCP server cards and 46 A2A agent cards in the same scan.

That does not mean only six x402 services exist. The Radar sample is a top-domain scan, not a census of every API host, marketplace, subdomain or machine-only service, and Cloudflare notes that x402 is only relevant to sites exposing paid resources. What it does show is a striking distribution gap: transaction infrastructure has moved faster than broad public-web exposure.

For sellers, that shifts the acquisition problem from merely implementing a 402 response to becoming readable, discoverable, callable and trustworthy to the agents that can actually pay.

Signal 03

Being listed is not the same as having buyers

A public 28 July full-registry survey of Coinbase Bazaar counted 14,865 listings. Under the author's deliberately simple repeat-demand rule — at least three unique payers and at least three calls per payer on average — 520 listings, or 3.5%, qualified. The survey counted 198 sellers with that repeat-use signal, and its top 10 sellers represented 89% of the qualifying calls.

x402stats reaches the same broad caution with a different method. Its August report classified 124 seller wallets as organic businesses out of 83,138 wallets that received anything, using a threshold of at least $100 revenue, at least three buyers and an average payment of at least $0.001. It reported 71% of its organic volume in the top 10 wallets. These datasets should not be combined: the filters, windows and units are different.

The useful overlap is directional. Listings, receiving wallets and payment events greatly outnumber sellers showing repeated buyer behavior under these independent filters. A machine-readable catalog improves reach, but it does not create a reason to buy.

The July Bazaar survey also found its largest repeat-use endpoints clustered around X/Twitter search, Tavily and Exa. That is one snapshot, not a universal category ranking, but it suggests a practical acquisition lesson: buyers were returning to services with a clear, already-understood data job. The x402 rail removed payment friction; it did not replace product-market fit.

Coinbase Bazaar full-registry survey →x402stats August report →

Signal 04

Discovery infrastructure is becoming a product in its own right

On 9 September, Circle launched a keyless x402 Discovery API that lets an agent search payable resources, filter by category, network and maximum price, inspect payment metadata and then decide whether to call. The launch is useful evidence that discovery remains an unsolved part of the buyer journey even as the payment rail matures.

Better catalogs can reduce search friction. They still cannot manufacture demand for an endpoint with an unclear job, weak evidence, poor reliability or an unattractive price. Seller acquisition therefore needs both machine discoverability and a product an agent can justify selecting.

zFinia'sx402 discovery guidetracks the current discovery surfaces and the checks a buyer should make before treating a directory result as callable inventory.

Signal 05

The product layer is becoming more credible

On 1 September, market-data provider Massive announced general availability of stock-market data for AI agents through x402 routes built with Coinbase. The important part is not another protocol demo: it is a familiar commercial data product being repackaged so software can purchase individual requests without creating an account, storing an API key or waiting for a human.

That is the adoption pattern worth watching: existing high-value digital services becoming directly purchasable by autonomous clients.

Signal 06

Trust and identity are joining the payment stack

On 10 September, Ant International, Mastercard and Visa announced work on interoperable standards for recognizing trusted purchasing agents. The direction matters for x402 builders even though it is a separate standards track: payment capability alone does not answer who authorized an agent, what it is allowed to buy or how a merchant should assess it.

Read zFinia's sourced breakdown of the emerging identity layer in theKnow Your Agent explainer.

What zFinia watches next

Independent repeat buyers

Repeated paid use is stronger evidence than raw request, listing or 402 volume.

Outcome-led demand

Clear, useful jobs should outperform endpoints that lead with the payment rail instead of the result.

Wallet-bearing discovery

Directories and routers matter when the agents searching them can actually settle a purchase.

Cross-protocol trust

Payment, discovery, delegated intent and agent identity are becoming separate but connected parts of the machine-commerce stack.

For buyers and builders

Inspect an actual machine-readable commercial surface

zFinia publishes its own paid machine services with the job, current catalog price, request example and contract separated from market-level claims. The catalog is useful as an implementation example; its existence is not presented as evidence of independent demand.

Explore the zFinia machine API →Machine-readable service catalog →

Find x402 APIs and services →Verify an x402 API before paying →What is Know Your Agent? →Machine Economy Daily →x402 protocol record →

Sources and methodology

Different x402 trackers use different chain coverage and classification rules. zFinia keeps each number attached to its source and does not blend incompatible totals into a synthetic market statistic. Public-web adoption, marketplace listings, receiving wallets and repeat-use filters are separate measurements from on-chain transaction activity.