Key facts
- Visa and Artemis report that infrastructure gaps are limiting the adoption of autonomous AI agents.
- Current payment systems are ill-equipped for the high-frequency, low-fee micropayments required by AI agents.
- New payment infrastructure is essential as AI agents have reached a capability threshold for autonomous transactions.
- AI-enabled microbusinesses could generate an additional $262 billion in stablecoin volume by 2033.
- Coinbase's x402 payment protocol has processed over 109 million transactions since May 2025.
- A unified payment framework could integrate stablecoins and traditional card transactions for AI agent commerce.
A joint report by Visa and Artemis has identified significant infrastructure challenges that are impeding the widespread commercial adoption of autonomous AI agents. The current global card payment systems, designed for human-initiated, low-frequency transactions, are struggling to accommodate the high-frequency, near-zero fee micropayments that AI agents require for autonomous commerce.
The report highlights that AI agents have recently crossed a critical capability threshold, enabling them to autonomously discover unfamiliar APIs, evaluate prices, and execute payments. This fundamental shift in commerce is being constrained by existing infrastructure gaps, according to the findings.
Projections suggest a substantial impact, with AI-enabled microbusinesses potentially driving an additional $262 billion in stablecoin volume by 2033, assuming a 33% adoption rate. This growth is predicated on AI-native payment solutions.
Emerging agentic payment standards are showing early signs of adoption. Coinbase's x402 payment protocol, launched in May 2025, has processed $15 million in adjusted volume across over 109 million transactions. The protocol saw a significant surge in October 2025, with monthly transactions increasing from 40,000 to 3.8 million.
Visa and Artemis propose that a unified machine-payments framework could support both stablecoin and traditional card transactions. This convergence would allow card networks to integrate into agentic payment flows, with stablecoins potentially serving machine-native micropayments and cards facilitating proxy purchases within existing merchant networks.
Tempo's Machine Payment Protocol, which spans both on-chain crypto and fiat payments, exemplifies this trend. Visa's own Card Specification SDK is designed to extend such protocols into card-based agent commerce. Both Visa's crypto division and Stripe-backed Tempo launched AI tools in March 2025 to facilitate same-day payments and streamline AI actor transactions.