The x402 payment protocol â" which enables AI agents to pay each other for services rendered using USDC and other stablecoins on a per-call basis rather than through subscription or monthly commitment â" has crossed a meaningful adoption threshold. The protocol is now processing over 2 million AI-to-AI payment transactions per day, with transaction volume growing at roughly 35% month-over-month.

The per-call model is the structural innovation that makes AI-to-AI commerce viable at scale. Traditional API pricing â" monthly commitments, rate limits, tiered pricing â" works for human customers who have predictable usage patterns. AI agents have bursty, unpredictable, and often very short-duration workloads that don't fit subscription models. Paying per call means agents can access specialized capabilities when they need them without committing to infrastructure they won't use.

Why Micropayments Enable Agent Economies

The economics of specialized AI agents create strong incentives for per-call payment models. A domain-specific agent â" say, one that specializes in legal document review â" shouldn't be running constantly. It should be invoked when another agent or an end user needs legal review. The per-call model allows the legal agent to charge for exactly what it delivers, and allows the calling agent to pay only for the specific capability it needs at the moment it needs it.

The x402 protocol's approach to payment authentication â" using cryptographic proofs rather than API keys â" also reduces the friction of cross-organizational agent-to-agent transactions. An agent representing one organization can prove its authorization to pay without requiring a direct billing relationship with the receiving organization. This is infrastructure that enables genuinely open agent marketplaces, which is why the payment volume growth rate has outpaced the growth of any single agent marketplace.