I analyzed all 14,865 listings in Coinbase's x402 agent-payment registry (github.com)

A recent analysis of Coinbase's x402 agent-payment registry reveals that, out of 14,865 API listings, only 3.5% show signs of genuine repeat usage, highlighting a significant discrepancy in the perceived demand versus actual usage. The x402 Bazaar serves as a public discovery platform for machine-payable HTTP endpoints, but with total organic volume at about 260,000 calls per month—generating only $3,000 to $6,000 in monthly gross merchandise value—this catalog is not a runaway success. Notably, 89% of this organic volume is concentrated in just ten sellers, primarily established brands rather than new entrants. This finding underscores critical implications for the AI/ML community: the metrics currently used to assess and rank API endpoints, which focus on raw settled volume and payer counts, may misrepresent real demand. Many listings exhibit patterns typical of one-off registration or incentive farming rather than ongoing usage, diminishing their value in genuine application development. The study is vital for developers and agents as it emphasizes the need for more accurate metrics to gauge API efficacy, encouraging a shift from inflated raw numbers to a focus on meaningful user engagement and consistent repeat transactions. The openness of the method for analysis allows others to verify these trends, ensuring that the x402 ecosystem can evolve with better, more reliable data.