AI Micropayments Point to Crypto’s Role in Agentic Commerce

AI agents are already using cryptocurrency to buy digital services, and the interesting part is not the spectacle but the economics: tiny, frequent payments that card rails were never designed to handle efficiently. For PSPs and acquirers, that makes agentic commerce less of a theory piece and more of a potential flow-shaping problem.

1. A Chainalysis study found more than 138 million transactions through x402 on Base, Coinbase’s layer-2 blockchain network. Per Decrypt, the activity reflects developers giving agents budgets to automatically buy computing power, live web data and coding tools, with payments becoming part of how the agents complete assignments.
2. The transaction sizes explain why this matters. More than 87% of payments were between a hundredth of a cent and 10 cents. Chainalysis looked at micropayments over the 16 months ending September and extended its review because the technology is new. The practical point is straightforward: these are the kinds of purchases where conventional card processing can become uneconomic.
3. In agentic commerce, the model is that software assembles what it needs while it works. An agent could buy fresh market data, obtain a model response and pay for a code review separately, rather than bundling all of that into one human-style checkout. That opens a monetization path for sellers of narrowly defined digital services and information products.
4. Stablecoins are central to that setup, according to the report, because their payment economics can handle amounts that card rails would swallow in fees. The crypto angle here is not speculation; it is settlement for useful services, with demand tied directly to work software is performing.
5. The supporting stack is moving in the same direction. 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. 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.”

For high-risk payment businesses, the takeaway is that automated purchasing is starting to look less like an edge case and more like a new traffic class: small tickets, high frequency, and service consumption driven by software rather than a person at checkout. That is exactly the sort of flow where payment economics decide whether the business works.