Payments Giants Back x402 Standard for AI Agent Transactions

Visa, Mastercard, Stripe, Google and 36 others formed a Linux Foundation body to steward x402, an open standard letting AI agents pay for services directly over HTTP.

By Daniel Mercer Edited by Maria Konash Published:
Payments Giants Back x402 Standard for AI Agent Transactions
The Linux Foundation launched the x402 Foundation with 40 members, including Visa, Mastercard and Stripe, to standardize AI agent payments. Image: x402 Foundation

The Linux Foundation announced the operational launch of the x402 Foundation, a new open-governance body backed by 40 organizations to steward x402, an open standard that lets AI agents, APIs and applications send and receive payments directly over the web.

Coinbase, which created the protocol, has formally contributed it to the foundation, moving it from a company project to a shared, vendor-neutral industry asset governed under the same model used for the Linux kernel and Kubernetes. The breadth of the membership is the headline: it spans traditional finance, cloud infrastructure and crypto, and includes every major US card network alongside big technology firms.

The premier members, who hold board seats, include Visa, Mastercard, American Express, Stripe, Adyen, Fiserv, AWS, Google, Shopify, Cloudflare, Circle, Ripple and the Solana and Stellar foundations. The technical premise is deceptively simple. x402 revives the long-dormant HTTP 402 “Payment Required” status code: when an agent requests a paid resource, the server replies with a 402 and a price, the agent signs a payment, usually a stablecoin transfer such as USDC, and repeats the request with proof attached, receiving the resource in seconds.

The exchange needs no account, no API key and no prior relationship. That matters because an autonomous agent cannot open a bank account, pass a credit check or sign a contract, but it can sign a transaction.

The launch reflects a broader industry conviction that software agents are becoming economic actors that need to pay for the data, compute and services they consume. Google has wired x402 into its own agent payments protocol, Cloudflare ships it in its agent toolkit and recently built a monetization gateway on it, and Coinbase has connected it to a new product giving approved agents access to trading and payment workflows. The founding companies frame an open standard as essential to prevent any single firm from controlling the payment layer of an emerging agent economy.

Why the Card Networks Showed Up

The most telling detail is that Visa, Mastercard and American Express are paying members of a standard that settles primarily in stablecoins, a technology once seen as a threat to their business. Their participation is defensive and strategic: rather than let a crypto-native rail bypass them entirely, the card giants are positioning to remain relevant in machine-to-machine commerce, with Mastercard promoting its “Agent Pay for Machines” and Visa arguing that future commerce will depend on interoperability across networks and payment types.

The mix of incumbents and crypto players in one body signals that no one wants to be left out of writing the rules for how agents transact, even as it papers over genuine competition beneath the shared framework, with Solana, Ripple and Circle each pushing their own settlement rails inside it.

The Reality Check

For all the marquee names, x402 remains tiny and unproven at scale. The network processed about 75 million transactions worth roughly $24 million over the past 30 days, an average of well under a dollar each, a rounding error next to the trillions moved by traditional payment systems.

The model also raises hard questions the foundation has yet to fully answer: whether consumers will trust autonomous software with access to money, how fraud, disputes and refunds work when there is no human in the loop, and how stablecoin settlement squares with evolving regulation.

Skeptics were quick to note the discomfort of handing an AI agent a wallet. The standard solves a real technical gap, but whether an agent economy large enough to justify it actually materializes is the bet every member is making, and it is still far from settled.

AI & Machine Learning, News