AI agents made 138 million crypto payments in a year, with over 87% in micropayments

AI agents made 138 million crypto payments in a year, with over 87% in micropayments
AI agents make 138 million crypto payments via x402

​The number of cryptocurrency payments made through Coinbase's x402 protocol exceeded 138 million by the third quarter of 2026. According to a Chainalysis study, more than 87% of transactions ranged from $0.0001 to $0.10, indicating growing demand for micropayments among automated services.

The findings were published in Chainalysis' report on cryptocurrency adoption in North America. Researchers examined the use of x402, a protocol that allows software applications to pay for digital services autonomously.

In mid-2025, the number of payments processed through x402 on the Base network stood at approximately 1,500. By the third quarter of 2026, the cumulative total had risen to more than 138 million transactions.

U.S. accounts for majority of AI agent payments

Developers use x402 to pay for computing resources, data access and programming tools. The small size of these transactions makes traditional card payments less economically attractive.

According to Chainalysis, wallets linked to the United States accounted for 88.1% of geographically identified payments for AI model usage. China's share was 0.1%, although Chinese wallets dominated spending on security checks and code analysis.

Analysts emphasized that the statistics cover only transactions for which a geographic location could be established. They do not represent the entire global market for AI agent payments.

Visa, Mastercard, Stripe, Coinbase and Google are also developing infrastructure for such transactions. These companies are creating tools that allow software applications to make payments without direct human involvement.

BlackRock also sees AI agents as a potential source of demand for cryptocurrency payments. Will Su, the company's head of digital asset research, highlighted the potential of round-the-clock transactions involving small amounts.

Not all automated payments involve autonomous AI agents

TRM researchers previously warned that some transactions attributed to AI agents may actually be executed by conventional software scripts. In addition, certain transfers may occur between related parties without representing genuine purchases of services.

As a result, transaction counts alone cannot determine the scale of independent economic activity carried out by artificial intelligence.

Meanwhile, Robinhood CEO Vlad Tenev described AI agents as a promising tool for retail investors. He believes such systems could perform tasks and execute trades even when users are not actively interacting with the platform.

Tenev suggested that AI agents could eventually handle a significant share of retail trading activity. However, he stressed the need to maintain human oversight of automated decisions.

Earlier, we reported that Anthropic suspended internal testing of AI agents with internet access due to security concerns and the risk of losing control over their actions.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.