Mastercard Tests AI Agent Payments for Digital Services Using Stablecoins and Fiat Currency

Mastercard demonstrated its Agent Pay for Machines (AP4M) payment platform, which allows AI agents to pay for digital services independently within a set budget. During the demonstration, an AI agent made 9 payments to 2 independent service providers.
At the Cloudflare Connect conference, Mastercard presented the results of testing Agent Pay for Machines, a technology designed to automate payments between AI agents and digital service providers. The solution allows agents to purchase data, computing power, and other digital resources without constant human involvement while allowing companies to maintain control over spending and agent permissions.
Mastercard first announced Agent Pay for Machines in June 2026. The solution is based on the open x402 protocol and enables AI agents to automatically pay for digital services within predefined spending limits.
During the test, marketing agency Hype tasked an AI agent with preparing an advertising campaign for a protein bar. Within the approved budget, the agent purchased social media research from AIsa and commissioned 8 advertising images through Cloudflare AI Gateway. In total, the agent made 9 payments to 2 providers without needing to open separate accounts or prepare payment documents for each transaction.
Mastercard used technical solutions from several partners to process the payments:
Coinbase Developer Platform provided digital wallet infrastructure to fund the agent;
BVNK managed the agent’s payment wallet and handled settlements through Mastercard Global Money Platform, which supports fiat currencies and stablecoins;
Base facilitated final blockchain settlement.
The payment model allows service providers to receive payments in their preferred currency or digital asset regardless of the payment channel used.
Following the tests with Cloudflare, Ant International, AIsa, Base, BVNK, Coinbase, and Hype, Mastercard also launched its Agent Pay for Machines Early Access program. Developers were given access to APIs, testing tools, and documentation for building automated payment services between AI agents.
For more information on how M2M payments between AI agents work, read CP Media’s dedicated article.
