AI Micropayments Point to Crypto’s Role in Agentic Commerce

 | 

A Chainalysis study found more than 138 million transactions through x402 on Base, Coinbase’s layer-2 blockchain network. As described by Decrypt, the activity reflects developers giving agents budgets to acquire computing power, live web data and coding tools automatically, making payments part of how the agents complete assignments.

The size of those purchases explains the opportunity. More than 87% of payments were between a hundredth of a cent and 10 cents. Chainalysis examined micropayments over the 16 months ending September, extending its review because the technology is new. Such payments could support businesses selling individual pieces of information or narrowly defined services on demand.

For agentic commerce, the implication is that software could assemble the resources needed for a task as it works. An agent might purchase fresh market data, obtain a model response and buy a code review, paying separately for each input. That would give service providers another way to monetize digital products and let customers purchase precisely what a task requires.

Stablecoins are central to that prospect, per the report. Their payment economics can accommodate amounts that would be overwhelmed by conventional card processing costs. For crypto’s real-world use, the significance lies in settling purchases of useful services, with demand tied to the work software performs. The opportunity could expand as businesses delegate more research, development and operational tasks to agents.

The supporting infrastructure is broadening. Decrypt reported in June that Visa, Mastercard and Stripe have moved to support agent payments, while Coinbase has introduced transaction tools and business acceptance capabilities. Google has also unveiled an agent-payments protocol. These developments suggest providers across the payments industry expect automated purchasing to become commercially significant.

BlackRock’s head of digital assets research, Will Su, told Chainalysis that agents could create “meaningful demand for frequent, low-value payments around the clock.” That pattern would create a different customer profile for payment providers, with software generating continuous streams of purchases.

The report also found signs of repeat use. Among qualifying wallets, the proportion receiving a second funding within 30 days rose from 70.6% for those first funded in the third quarter of 2025 to 89.3% for the second quarter of 2026. The analysis excluded wallets that did not send payments on at least two separate days. That selection matters when assessing how broadly retention extends.

Geographic findings require similar care. American-linked wallets accounted for 88.1% of geographically identifiable model-inference payments, compared with 0.1% for China. Chinese-linked wallets were prominent in security and code-analysis purchases. These figures describe identifiable x402 activity, rather than the global market for AI services.

The largest qualification concerns who is actually transacting. Decrypt cited earlier TRM research suggesting that scripts and self-dealing may account for much apparent agent activity. High transaction counts therefore cannot establish either widespread autonomous adoption or substantial commercial value.

For businesses considering deployment, those limitations make verification essential: whether payments purchase useful services, whether customers return and whether spending remains within authorized budgets. The findings support a credible path for crypto in agentic commerce. Its durability will depend on repeat economic utility and accountable spending as automated purchasing grows.