This week the phrase "machine-native economy" is everywhere. BlackRock put its name on it, and every feed has redrawn the same diagram: a human paying a merchant today, an agent paying an API tomorrow. The rails behind that picture are real. x402 lets a web service charge per request. ERC-8004 gives agents an on-chain identity. Stablecoins settle in seconds.
What almost nobody publishes is a careful count of what agents actually do with money. So we built one. For every AI agent registered on Base, we follow the wallets bound to it and record every dollar that leaves them. Here is what that count shows.
Today: what the counted money actually is
As of 2026-11-10 18:12 UTC: 31,141 agents have been seen paying someone, out of 1,177,680 registrations on Base (1 in 38). Counted spending is $57,457,338, with $626,768,311 of pre-registration history and $181,009,823 of routing hops excluded beside it. Every figure in this box is live; the dated sections below are kept as they were measured.
A larger total is only worth publishing with its destinations named, so we identified the addresses that receive it, from verified contract source and from the transaction receipts. Of counted agent spending:
| Where it lands | Share of counted spending | USD | What the evidence is |
|---|---|---|---|
| Cash parked at yield through lending adapters | 45.0% | $25,880,123 | The largest single recipient is a verified contract named Adapter whose own source says it wraps Aave v3 and ERC-4626 vault deposits and hardcodes the depositor as the receiver. The 82 agent wallets that use it are all Safe smart accounts with session-key modules; relayers and bundlers submit every move. Plus Morpho's adapter, Euler vaults, Aave, Compound, Fluid and Moonwell markets. |
| Bridged off Base | 17.8% | $10,206,277 | Across SpokePool, Relay, LI.FI, a CCTP wrapper, and three unverified routers whose every sampled receipt forwards the full amount to Across in the same transaction. |
| Work bought through escrow | 20.1% | $11,524,863 | TermixEscrow: order lifecycle with a dispute layer, in its verified source. The one large service market. |
| Swapped for other tokens | 3.8% | $2,204,861 | Uniswap v3 and Aerodrome pools, a fee-taking router and an aggregator: USDC in, another token back to the agent in the same transaction. |
| Other yield vaults | 1.3% | $767,777 | ERC-4626 vaults with verified names. |
| Not yet named | 12.0% | $6,873,436 | Mostly plain addresses. We do not guess. |
Receipts make the first row exact. $26,547,844 of counted spending, 46.2%, went into lending and vault positions that the transaction receipt shows were credited back to the paying wallet itself: the agents' own money, parked. What left their control on Base was $30,909,494, and that includes bridges, where the money may still be theirs on another chain. Definition →
So the money that registered agents control is mostly treasury management: lent out, moved to another chain, or exchanged. Purchases of work are 20.1% of counted spending, and x402 payments 1.71% of the count. It is also concentrated: the single largest wallet is 6.97% of the counted total, and the median payment is $0.05. Read the total beside those two numbers, never alone. Each label carries its evidence (a verified contract name, or the receipts showing where the money went next) at /api/partners and the definitions at /def; we do not name the companies behind the contracts, because a contract name is not evidence of who runs it. These rows are live and will move as the receipt check clears its queue.
Update, 27 September: the census
The figures further down are the 24 September measurement, kept as written. On 27 September the graph moved from a sample to a census: every ERC-8004 registration on Base is in it, every owner wallet is swept for stablecoin transfers every three hours, and each wallet's whole history is read rather than its newest fifty transfers. That day, agents observed paying rose from 443 to 2,387, and counted spending to $11,832,139, with $109,157,479 of pre-registration history and $29,910,781 of routing hops excluded beside it. Of the counted spending, $6,686,228 (56.5%) went into lending and vault positions the paying wallets still held. Values as of 27 September 2026, from our daily archive; today's are in the box above.
24 September: $4.86 million out, $757,000 spent
By 24 September 2026, $4,856,774 in USDC had left wallets bound to registered agents since the first registrations on Base in early February. It splits into three parts that add up to the dollar:
| What the money was | Amount | Share |
|---|---|---|
| An agent paying someone else | $757,082 | 15.6% |
| Routing hops: money passing through an agent wallet on its way somewhere else | $1,533,611 | 31.6% |
| Money the wallet moved before the agent was registered | $2,566,081 | 52.8% |
Only the first line is an agent spending money. The second is plumbing: the same dollars bounce through a wallet as part of a swap or a relay, and counting each hop as a purchase counts the same money several times. The third is history. Many agents are bound to wallets that existed long before the agent did, so counting that wallet's whole past as "agent activity" credits the agent with money it never touched.
A total that doesn't separate these three is more than six times too big. That's not a guess about other people's numbers; it is the gap in our own data before and after the exclusions. We publish both exclusions as their own figures, so anyone can check what we left out and why.
24 September: 95,520 names, 443 payers
Registrations are the other number people like to quote. On 24 September 2026, on our first sample, 95,520 AI agents were registered on Base and 443 of them had ever been seen paying anyone, 1 in 216. Today's count is in the box at the top.
A registration is a mint, not an agent. It costs almost nothing, and it can be done in bulk. 28.6% of all registrations were minted in bursts by single addresses, and the ten most active minters account for 35.1% of the whole registry. None of that is wrong, since platforms legitimately issue identities for their users. But a registry count measures how many names were issued, not how many agents are working.
We rank agents by evidence instead of by claims. An agent reaches level 7 (L7) on our ladder when a wallet bound to it is seen paying money out of its own control. L8 adds a funded wallet. L9 needs three independent counterparties. Most of the registry never gets past the name. How the ladder works →
The paying agents are concentrated too
Even among the 443 agents that paid by 24 September, the picture was narrower than the headline suggests.
One escrow contract, TermixEscrow, receives 25.5% of all counted agent spending: $193,091 from 214 agents. It's a work-escrow platform, where agents pay into orders that settle when the work is delivered. More striking: 189 of the 443 paying agents made their very first payment to that contract, inside eleven days in August.
That's real economic activity, but it is one platform onboarding many agents at once, not a crowd of independent buyers. When you remove that wave, new paying agents arrive at a steady six or seven a week. The agent economy on Base doesn't grow smoothly. It grows in platform-sized steps.
Reputation shows the same shape. Across 455,286 on-chain reputation events, ten raters wrote 51.7% of all the feedback. A signal that ten addresses can dominate isn't much of a signal.
Why the outside estimates disagree
We aren't the only ones asking whether agent payments are real. TRM Labs reportedly screened roughly half of x402 settlement volume out as self-payments and anomalies. Artemis called the x402 boom "still mostly a mirage" back in February. News outlets report tens of millions of dollars a month.
These numbers are hard to compare because they describe different populations under different rules. x402's public stats show around 10,000 unique payers a day on Base, but among agents that are actually registered on-chain, x402 was only 3.1% of payments by count on 24 September, and $23.40 by value. Most of the x402 traffic isn't coming from agents with an on-chain identity at all.
Nobody is lying here. The disagreement is a measurement problem: what counts, who counts, and whether the exclusions are published. That's the part we think matters most. Four nouns people add together →
What would make the number honest
Every figure we publish carries four things: what exactly is counted, on which chain, over what window, and as of when. Every metric has a definition id. And before anyone trusts the tables underneath, 37 automated checks run against them every three hours. Each check is a query that returns zero rows when the data is healthy. Every check, live →
We built those checks because every serious correction we have made so far was a data-shape error someone spotted by eye: routing counted as spending, a wallet's history before its agent existed counted as the agent's (that one inflated the headline by 46%), and a seller list that sat frozen at page one for eleven days while every run reported success. Each is now a named check, so the next one gets caught by a machine instead of a reader.
The honest summary
The rails for machine money exist, and they're spreading fast across chains and custodians. The money moving on them, measured carefully, is small: three quarters of a million dollars of real agent spending on Base, from a few hundred agents, a large share of it through one platform.
That isn't a discouraging number. It's the size of the thing that hasn't happened yet, and it's far more useful than a headline total built by adding things that can't be added. If agents do become a real payment channel, the first people to notice will be the ones counting carefully.
All figures: Agentic Finance Graph, Base mainnet, measured 24 September 2026. Live data and definitions: /api/state. Outside figures as reported by Decrypt (TRM Labs), Artemis on X and x402scan.