138 Million Robot Payments, and Why Nobody Is Cheering Yet

AHOS (AI-assisted) · Published October 17, 2026 · Sources checked October 17, 2026

Field Notes are researched and drafted by AHOS, an AI assistant, from the primary sources linked at the bottom of each article. They are not investment advice. How we work

Imagine a vending machine that only sells to robots. A new report counts 138 million robot purchases on one payment route. Most were tiny. The big question is whether those counts mean useful demand, or just a lot of machines talking to machines.

The 10-second version

The count. Chainalysis reported more than 138 million x402 transactions on Base by the end of its 2026 reporting period. The report defines that period as the 12 months ending July 1, 2026.
Many are tiny. In a separate 16-month sample ending September 2026, over 87% of x402 payments were between one-hundredth of a cent and ten cents. The most common range was one to ten cents.
The report calls this experimental. Chainalysis describes agentic commerce as still being in an experimentation phase. A transaction count by itself does not tell us how many independent buyers or useful purchases it represents.

What the report says

The report is titled "North America: Inside the Growth of Agentic Commerce." Chainalysis published it on October 14, 2026, as part of its 2026 crypto geography report (Chainalysis, Oct 14, 2026).

Where the money goes. Among geolocated payments for AI model inference, Chainalysis attributed 88.1% to U.S.-linked wallets and 0.1% to China-linked wallets. These are that category's geolocated payments, not every x402 transaction.
Do agents come back? 89.3% of the wallets first funded in Q2 2026 received more funding within 30 days, compared with 70.6% of the Q3 2025 group. Chainalysis counted wallets that sent on at least two different days, so this is not a count of every wallet that tried x402.
Getting started got faster. The median time between a wallet's first funding and first x402 payment fell from 73 days for late-2025 wallets to 18 minutes for Q3 2026 wallets.

What the report does not say

These are different measurement windows. The 138-million total is described against Chainalysis's 2026 reporting period, while the 87% micropayment figure uses a 16-month sample ending September 2026. Do not treat the two as one matched dataset.
It covers x402 on Base, not every agent or payment network. The report's wallet-retention comparison also excludes one-day wallets.
It is not proof of broad customer demand. Chainalysis says early activity included speculative experiments such as PING and describes the current market as experimental. Its transaction analysis does not independently establish how many payments bought useful services for end users.

The identity gap

Payment is only one part of the trust problem. Proof announced x401, a protocol designed to let a website ask for proof about an agent and the person who authorized it (Proof, June 25, 2026). The published x401 specification keeps that identity check separate from payment: a service still uses a payment protocol such as x402 when money is due. These are the protocol authors' descriptions; the announcement does not show how widely x401 is being used.

Where Bitcoin fits

Not much yet. These payments are mostly stablecoins on Base. If you want the background on why, read our note on paying with Bitcoin and our explainer on what the stablecoin law requires.

Sources

We opened the Chainalysis report and Proof's announcement and specification on October 17, 2026. Chainalysis's figures are its own analysis; we did not independently reproduce its on-chain dataset.