Mastercard, Coinbase and blockchain network XDC are building payment rails that let AI agents buy services on their own, without a person entering card details or clicking approve. The scale of the opportunity explains the interest: McKinsey estimates AI agents could mediate between $3 trillion and $5 trillion in global consumer commerce by 2030. If software is going to complete tasks on its own, it needs a way to pay for what it uses.
What XDC and x402 are building
XDC AI gives an AI agent the ability to find and pay for digital services on demand instead of going through a traditional human checkout. A user grants the agent access to a wallet and sets a spending limit. When the agent needs data from a paid API to finish a task, it makes the request and pays in USDC, with XDC covering the gas fee, so the agent does not need to hold XDC to complete the payment. XDC Tech, the US institutional arm of XDC Network, has integrated with Bridge, a Stripe company, connecting onchain payments to traditional financial infrastructure and giving developers tools to move between fiat currencies and stablecoins.
The mechanics rest on x402, an open protocol developed by Coinbase. It revives HTTP 402, an old web status code meaning «Payment Required» that existed for years without wide use. An agent requests a paid API, the server answers with a price instead of the data, and the agent's wallet pays automatically: no login, no card number, no human in the loop. XDC has built that flow into its own marketplace, where an agent can find an API, pay per use in USDC and stay within the limit its owner set. An agent researching a market might need several paid data sources, a specific API or a small amount of computing power, and pays only for what it needs instead of opening a new account or subscription each time.
Mastercard is preparing for the same shift. When it launched Agent Pay for Machines in June, the company described a future in which agents continuously buy services from each other, including payments worth fractions of a cent. Atul Khekade, co-founder of XDC Network, said every internet transaction so far has assumed a person is on the other end, and AI agents break that assumption completely. Adoption remains early, but the Linux Foundation now oversees the x402 Foundation, whose 40 members include Mastercard, Visa, Stripe, American Express, Google, AWS and Circle.
What this means for business
For companies that run paid APIs, data services or computing capacity, the change is practical: machine buyers can pay per call in stablecoins without a sales conversation, an account or a subscription. A small vendor can sell a single API request to an agent, while a large platform can meter thousands of transactions a day. Gartner expects agentic AI to be built into 33% of enterprise software applications by 2028, up from less than 1% in 2024, which turns agent traffic into a customer segment that needs pricing and billing built for it.
The limits matter as much as the promise. Fully autonomous shopping is still rare, and AI is far better at helping people compare options than at completing purchases on their own. The gap between x402's onchain activity and the valuation of its ecosystem shows that infrastructure and adoption rarely move at the same pace: the settlement rails, wallet permissions and integrations with players like Bridge are real, but the volume moving through them today is a fraction of what the surrounding numbers suggest. A business evaluating a provider should ask who holds the wallet keys, how spending limits are enforced, which stablecoin settles the payment, who covers gas fees and what happens when an agent pays for a service it cannot use.
The contest is splitting along two lines. Card networks are betting that trust and identity, the problem they spent decades solving for people, will matter just as much for machines. Networks built for stablecoin settlement, XDC among them, are betting the opposite: that speed and near-zero fees will define the market once agents transact at scale. The better approach will likely depend on the use case, and the infrastructure decisions being made now by Mastercard and XDC will shape how autonomous commerce works once it arrives.
A useful marker to watch is the transaction volume flowing through x402 and similar rails over the next year, rather than the number of members in the x402 Foundation or the valuations attached to the sector. If that volume grows alongside the 33% of enterprise applications Gartner expects by 2028, machine payments will have moved from experiment to routine cost line. If it stays flat while membership lists grow, the infrastructure is still ahead of the demand.
