Key Takeaways:
Mastercard's new payment protocol lets AI agents buy and sell services autonomously, with transactions as small as fractions of a cent.
Mastercard on June 10 launched Agent Pay for Machines, a payment protocol that lets AI agents authorize and settle transactions autonomously across its global network, with more than 30 partners including Coinbase, Stripe and Ripple.
"Agent Pay for Machines will create the conditions for a superbloom of AI business models," Jorn Lambert, Mastercard's chief product officer, said. "Machine payments can make it possible for services to be bought and sold among agents at fundamentally different scales than payments today."
The protocol handles transactions as small as fractions of a cent, executed continuously in the background of digital commerce. Settlement spans cards, bank accounts and stablecoins, with agent permissions logged on Polygon, Solana and Base blockchains. The system credentials each agent, enforces programmable spending limits, and supports multi-rail settlement across providers.
Lambert described AP4M as a five-year bet on machine-to-machine commerce rather than a near-term revenue driver. The launch puts Mastercard alongside Visa, Stripe and Google in the race to build payment rails for an economy where software, not people, initiates purchases — a market that could reshape how digital services are bought and sold.
How the protocol works
Agent Pay for Machines builds on Mastercard's Agent Pay program introduced in 2025, which defined how trusted AI agents make purchases for people. The new system targets automated, high-frequency payments between machines and services through four functions: credentialing each agent with Verifiable Intent identity verification, permissioning spending rules that are enforced programmatically, enabling agents to transact across systems, and settling payments across cards, accounts and stablecoins.
Mastercard chose to record agent permissions on public blockchains rather than in a private database, allowing multiple parties to independently verify whether an agent operates within its human owner's boundaries. The initial chains selected are Polygon, Solana and Base.
In one example, an entrepreneur instructs an AI agent to launch a flower shop's website. The agent buys a domain, hosting, images and checkout pages within a set budget, turning a single request into a chain of automated purchases across providers.
Who wins, who loses
The partner list spans payments infrastructure, crypto platforms and blockchain networks. Stripe, Adyen and Checkout.com bring merchant processing capabilities. Coinbase, OKX and Ripple provide digital asset settlement rails. Cloudflare contributes AI agent security infrastructure. Polygon, Solana Foundation and Base offer blockchain-based permission logging.
The breadth of participation signals that the industry sees machine-to-machine payments as a necessary layer for agentic commerce, even if the timeline remains uncertain. Lambert told Fortune he does not expect AP4M to be "a huge revenue driver for Mastercard next year," but sees it becoming "a meaningful new addressable market" over the next five years.
Visa and Stripe have released their own agent payment tools over the past year, though transaction volumes remain a small fraction of total commercial flow. Google has also entered the space with agent-focused payment infrastructure. The competition suggests that payment networks view the AI agent economy as a potential new revenue stream that could eventually process billions of microtransactions annually.
Investment angle
Mastercard shares trade at about 35 times forward earnings, a premium reflecting its position as a toll collector on global consumer spending. AP4M represents a potential expansion of that toll-collection model into machine-driven commerce, though Lambert's five-year timeline suggests any revenue contribution is distant. For crypto platforms like Coinbase and Ripple, the protocol validates stablecoins and blockchain rails as infrastructure for the emerging AI agent economy, potentially driving long-term demand for settlement tokens and payment platforms. Polygon and Solana gain a use case for their blockchains as permission registries for autonomous agents.
This article is for informational purposes only and does not constitute investment advice.
