The payment ecosystem for AI agents is beginning to take clear shape. A joint analysis by experts from Visa and the blockchain analytics company Artemis has revealed a fundamental division: traditional cards and stablecoins will serve fundamentally different segments of automated transactions.
Specialists identified two key types of operations. The first is purchases made on behalf of a human: booking tickets, managing subscriptions, ordering goods. These transactions are essentially no different from regular online shopping and can be processed through standard card infrastructure. The second type is micropayments between programs: frequent transfers of amounts under $1 for individual API requests, access to computing power, or the use of software tools.
This is where the key advantage of stablecoins lies. The fixed fees of card networks make them unprofitable for microtransactions. At the same time, on the Base and Solana networks, the cost of conducting an operation is fractions of a cent. The report examines two open protocols: x402 (activity concentrated on Base, Solana, and Polygon) and the Machine Payment Protocol (MPP), which supports settlements both through blockchains and traditional systems.
The most likely model for the future appears to be a hybrid approach: cards are used for authorization and large purchases on behalf of a human, while stablecoins are used for settlements between agents. Both instruments can be applied at different stages of completing a single task.
Real figures, however, are still modest. According to Artemis data as of April 21, after filtering out fictitious and test operations, approximately $15 million passed through x402 across 109.6 million transactions. For MPP, the figures were about $25,000 and 115,000 transfers, respectively. Without filtering, the numbers look significantly more impressive: $135.7 million and 178.3 million operations for x402.
The main unresolved issue remains trust. Who bears responsibility for an erroneous or unauthorized purchase: the agent owner, the platform operator, the model developer, or the seller? Existing payment reversal mechanisms are designed for human operations and do not work with thousands of automatic transactions per hour. An additional risk is malicious instructions that could redirect a payment or initiate an unplanned purchase.
My expert assessment: Stablecoins are indeed becoming an ideal tool for machine-to-machine settlements, but mass adoption is hindered not by technology, but by the legal framework. Until a clear system of responsibility and transaction dispute mechanisms is established, AI agents will remain a niche experiment. However, the trend is obvious: the future of payments lies in hybrid models, where each tool occupies its own niche.