TRM research finds that AI agents are not yet a significant portion of x402 protocol spending, despite $52.7M in total settlements.
AI & Agents ·
Blockchain intelligence firm TRM Labs examined $52.7 million across 198.9 million settlements on the x402 protocol, finding that AI agents account for only a small fraction of activity. After filtering out self-payments and anomalous flows, the research estimated that 0.6%–7.5% of the remaining $25.62 million in likely commerce came from AI agents, suggesting the technology has not yet become a significant driver of protocol spending despite growing industry investment in agent payment infrastructure.
The research reveals fundamental measurement challenges. TRM's models cannot conclusively distinguish AI agents from scripts, scheduled jobs, load tests, and self-dealing arrangements—all of which produce identical blockchain records. The firm noted that many current agents may be single-purpose systems that repeatedly pay one service, which would be misclassified as scripts under its methodology, potentially understating true agent commerce.
Beyond quantification gaps, TRM identified infrastructure shortfalls in accountability and oversight. On-chain agent registries exist but participation remains voluntary and underutilized. The report calls for improved registration standards, counterparty reputation systems agents can verify, and compliance monitoring designed for high volumes of small payments rather than traditional value-based thresholds. What technical infrastructure exists to support agent payments, the research suggests, outpaces the governance frameworks needed to manage them.