Stablecoins are becoming the "circulatory system" for micropayments of AI agents: Visa analysis
Experts from Visa and the blockchain analytics firm Artemis conducted an in-depth analysis of payment flows in the field of artificial intelligence. Their conclusion is clear: traditional cards and stablecoins will occupy fundamentally different niches in servicing AI agent transactions.
The study's authors divided operations into two distinct classes. The first is purchases made on behalf of a human: booking tickets, managing subscriptions, ordering goods. Essentially, this is classic e-commerce, where payment cards are convenient and familiar. The second class is microtransactions between programs: paying for individual API requests, data access, computing power, or software tools. The amounts here rarely exceed $1.
This is precisely where, in my assessment, the key breakthrough lies. The fixed fees of the card infrastructure make it economically unviable for such small payments — they simply "eat up" the entire amount. Meanwhile, on the Base and Solana networks, fees amount to fractions of a cent. The report examines two open protocols in detail: x402 (active on Base, Solana, and Polygon) and the Machine Payment Protocol (MPP), which supports both blockchains and traditional systems.
According to Artemis data as of April 21, after filtering out test and dummy transactions, approximately $15 million passed through x402 across 109.6 million transactions. MPP shows more modest figures: $25,000 and 115,000 transfers. For comparison, the raw data was several times higher: $135.7 million and 178.3 million operations via x402. The difference is the result of cleaning out clearly artificial activity. This indicates that the real market is only just forming, but its potential is enormous.
The authors propose a hybrid model: cards for authorization and purchases made on behalf of humans, stablecoins for settlements between programs. Both methods can be applied at different stages of a single task.
The main unresolved issue is trust. Who is responsible for an erroneous or unauthorized purchase: the agent owner, the platform operator, the model developer, or the seller? Existing chargeback rules are designed for humans. There is currently no mechanism for disputing thousands of automatic transactions per hour. An additional risk is malicious instructions capable of redirecting a payment or forcing an agent to make an unplanned purchase.
My expertise: the market for micropayments for AI agents is a "blue ocean" for stablecoins. Until the trust infrastructure and transaction dispute standards are developed, mass adoption will remain limited. But it is precisely blockchain solutions that offer the most elegant way to solve the "micro-fee" problem — and Visa understands this perfectly.