AI agents outgrow card networks, boosting stablecoin micropayments
A joint Visa and Artemis report warns that traditional card networks cannot process the high-frequency micropayments required by AI agents, opening a $262 billion market opportunity for stablecoin infrastructure.
Traditional card payment infrastructure is fundamentally unsuited for the emerging economy of autonomous AI agents, according to a joint report published Wednesday by Visa and Artemis. The current system was built for human commerce, characterized by low-frequency transactions, and lacks the near-zero fees and rapid settlement required to make machine-driven micropayments commercially viable.
This infrastructure bottleneck arrives just as AI agents crossed a critical capability threshold in mid-2025. These systems can now independently discover unfamiliar APIs, evaluate pricing, and execute payments. However, without adequate payment rails, mainstream commercial adoption remains stalled. The gap points to a significant shift toward stablecoins for machine-to-machine transactions. Australian exchange Swyftx estimates that AI-enabled microbusinesses could generate an additional $262 billion in stablecoin volume by 2033, assuming a 33% adoption rate.
Early data suggests substantial latent demand for these specialized rails. Coinbase’s x402 payment protocol, a standard designed for agentic payments, has processed $15 million in adjusted volume across more than 109 million adjusted transactions since launching in May 2025. Usage accelerated sharply in October, surging from 40,000 to 3.8 million monthly transactions, culminating in 38 million transactions processed that month alone.
Incumbents are moving to adapt rather than be displaced. Visa and Artemis project that a single machine-payments framework will eventually support both traditional card and stablecoin flows. “The trajectory points toward convergence rather than competition: cards for proxy purchases inside existing merchant networks, stablecoins for machine-native micropayments, and hybrid flows where both are used within the same workflow,” the report stated.
Tempo’s Machine Payment Protocol already bridges onchain crypto and fiat payments using shared tokens, while Visa’s Card Specification SDK aims to integrate card networks into these agent-driven workflows. Both Visa’s crypto division and Stripe-backed Tempo introduced dedicated AI payment tools in March, with Visa enabling same-day agent payments. For investors and financial institutions, the report signals that the future of commercial payments infrastructure will depend heavily on integrating crypto-native settlement layers to handle the sheer volume of automated machine transactions.
