Experts from payment giant Visa, together with blockchain analytics company Artemis, conducted an in-depth analysis of the future of payments for AI agents. Their conclusion is clear: traditional cards and stablecoins will occupy fundamentally different niches in this ecosystem.
The authors divided all AI agent transactions into two categories. The first is purchases made on direct human instruction: booking tickets, managing subscriptions, ordering goods. These operations are essentially similar to conventional online commerce and can be serviced through classic payment cards.
The second category is high-frequency micropayments, where the amount of each transaction is less than one dollar. This includes paying for individual API requests, accessing computing power, purchasing data, or using software tools. It is here, according to analysts, that stablecoins demonstrate their undeniable advantage.
Fixed fees charged by card networks make their infrastructure economically unviable for such microtransactions. At the same time, on Base and Solana networks, transaction fees amount to fractions of a cent, which is ideal for automatic settlements between programs.
Two Protocols on the Stage
The report examines in detail two open payment protocols: x402 (actively used on Base, Solana, and Polygon) and the Machine Payment Protocol (MPP). The latter supports settlements both through blockchains and traditional payment systems, making it a hybrid solution.
Artemis data as of April 21 shows that through the x402 protocol, after filtering out fake and test transactions, approximately $15 million was processed across 109.6 million transactions. For MPP, these figures are significantly more modest: about $25,000 and 115,000 transfers. The original unfiltered data was notably higher — $135.7 million and 178.3 million transactions via x402, confirming the scale of the "junk" transaction problem in this area.
Key Issue — Trust
The authors of the study identify trust as the main unresolved issue. It remains unclear who should bear responsibility for an erroneous or unauthorized purchase made by an AI agent: the agent's owner, the platform operator, the model developer, or the seller.
Existing chargeback mechanisms are designed for transactions made by humans. An established mechanism for disputing thousands of automatic transactions per hour does not yet exist. Additional risk is posed by malicious instructions capable of redirecting a payment or forcing an AI agent to make an unplanned purchase.
The authors suggest the formation of a hybrid model: cards will be used for authorization and purchases on behalf of humans, while stablecoins will be used for settlements between programs. Both payment methods may be applied at different stages of a single task.
My expert assessment: This analysis clearly indicates that stablecoins are becoming not just a speculative tool, but a fundamental infrastructure for the AI agent economy. However, without solving the problem of legal liability for algorithm actions, the mass adoption of such payments will remain in question. The market awaits not only technological but also regulatory innovations.