Competing in the AI payment track, traditional card organizations versus Coinbase

Article by: Zennon Kapron

Translated by: Chopper, Foresight News

As AI intelligences increasingly undertake various commercial transactions, a battle over the underlying payment channels has already begun.

Currently, there are two major incompatible technical approaches for enabling AI intelligences to achieve autonomous consumption: one where software programs make payments on behalf of users, and the final settlement is completed through which channel. One camp relies on tokenized bank card credentials controlled by Visa and Mastercard to build the payment link; the other, led by Coinbase, uses open internet protocols and stablecoins to settle transactions. The superficial focus of AI e-commerce is on shopping assistant applications, but the core contest behind the scenes is who can dominate the next-generation payment system.

Two major payment channels, suited for different application scenarios

Traditional card organizations are taking the lead with swift actions. Mastercard launched its Agent Pay service in April 2025, built on its self-developed agent token system. This tokenization technology was originally used for contactless payments and quick payments linked to bank cards, but now it has been expanded to allow verified AI intelligences to make transactions within user-authorized scopes.

At launch, the service gathered many industry partners, with a clear strategic intent: partners include Microsoft, IBM’s WatsonX intelligent orchestration platform, as well as payment service providers Braintree and Checkout.com. One day later, Visa also launched its Intelligent Commerce service, opening its payment network to AI developers, with the core being AI-compatible bank cards. This scheme replaces original card numbers with tokenized credentials to prove user authorization for specific AI intelligences and delineate transaction permissions. Visa also engaged top AI companies, including Anthropic, OpenAI, Perplexity, Mistral, and Samsung.

Both card organizations’ solutions keep transactions within the traditional bank card payment model that has been used for decades. AI intelligences are new players, but behind them still run the traditional payment channels that have served global commerce for over half a century.

The stablecoin camp, on the other hand, has adopted a completely different architecture. Coinbase launched the x402 protocol in May 2025, reactivating the long-shelved HTTP 402 “Payment Required” status code, enabling direct internet transactions using USDC stablecoins for settlement. The process is as follows: the client requests access to a resource, the server responds with a payment instruction; the client attaches signed stablecoin payment information in the request header; once the on-chain transaction is confirmed, the resource can be accessed normally. This entire process requires no account registration, no bank card binding, and incurs no bank transaction fees.

This scheme is specifically designed for machine-to-machine transactions. AI intelligences may need to pay for API calls, data streams, or interfacing with other intelligences, completing thousands of small payments. If these transactions used traditional bank card channels, the costs would be completely prohibitive.

Each approach has its strengths. Bank card channels excel in personal retail consumption, where high standards for chargebacks, fraud prevention, and dispute resolution are required; stablecoin channels are highly advantageous for high-frequency, small, cross-border machine transactions, where traditional bank fees and settlement times would be entirely unsuitable. The core of the contest is which scenario will become the mainstream for AI business transactions.

A major challenge faced by both approaches is identity verification. When a software program initiates a payment, merchants need to confirm that the operator is a legitimate AI authorized by the real user, not a malicious bot using stolen credentials; simultaneously, users need mechanisms to revoke transactions initiated by AI intelligences in case of errors.

Visa states that AI access traffic on U.S. retail websites has surged 47-fold, and has partnered with cloud service provider Cloudflare to launch a Trusted Agent Protocol to distinguish legitimate AI programs from malicious crawlers. This is a structural advantage of traditional card organizations: their fifty-year-old risk control scoring systems, chargeback rules, and dispute handling mechanisms are well-equipped to handle issues like AI intelligences purchasing the wrong items. Once a stablecoin transaction is on-chain, it is permanent and cannot be rolled back, and currently, there are no solutions within the native system for this.

In the future, the key to the personal market’s direction may not be which payment channel charges lower fees, but rather who can solve the challenges of AI identity verification and transaction dispute resolution.

Card organizations’ dual-track layout, fully betting on both major routes

An intriguing signal is that Visa and Mastercard are not solely committed to their own channels but are simultaneously deploying in the stablecoin arena.

By April 2026, Visa’s stablecoin settlement transactions reached an annualized volume of $7 billion, a 50% increase month-over-month; the company added five new supporting public blockchains, bringing the total to nine, and has implemented over 130 “stablecoin + bank card” linkage projects across more than 50 countries worldwide. In October 2025, Visa further expanded, partnering with Cloudflare to launch the Trusted Agent Protocol, helping merchants identify legitimate agents versus malicious programs, and announced cooperation with Coinbase to enable interoperability between their network and the x402 protocol. The seemingly competitive bank card system and stablecoin protocols are now building bridges for interoperability.

Mastercard has adopted a similar dual-track strategy. In March 2026, Mastercard announced it would acquire the stablecoin platform BVNK for up to $1.8 billion. Prior to that, its agent payment services had expanded into Latin America and the Caribbean, completing local card issuer adaptations in early 2026.

It’s clear that the core strategy of these two traditional card organizations is no longer solely to defend their bank card channels but to become the fee gateways for all payment links, whether through their own channels or stablecoin channels. This layout indicates their judgment: if the industry ultimately adopts bank cards as the main AI payment method, they won’t need to make large investments in acquiring stablecoin infrastructure.

Application scenario differentiation

From the products already launched, the application boundaries of the two approaches are quite clear.

Mainstream products aimed at ordinary consumers mostly use bank card channels. For example, the ChatGPT “one-click checkout” feature launched in September 2025, jointly developed by OpenAI and payment service provider Stripe, relies on shared payment tokens to settle bank card payments. These tokens are limited to designated merchants and shopping orders, initially integrated with Etsy merchants, and later expanded to over a million Shopify stores. Amazon’s “Proxy Ordering” feature allows AI intelligences to purchase on third-party sites, automatically filling in the user’s linked bank card during checkout.

Personal consumer AI shopping services generally prefer bank cards because this system has mature fraud prevention tools, a well-established merchant network, and long-term user trust.

Meanwhile, stablecoin channels dominate the machine transaction market. Amazon integrated the x402 protocol into its Bedrock AI core payment service, settling via Coinbase’s Base blockchain, with each transaction taking about 200 milliseconds and costing less than one cent. Stripe also joined as a payment partner. Coinbase data shows that in the first year after x402’s launch, over 169 million payments were processed, involving 590k buyers and 100k sellers.

These transactions are not typical online shopping for consumers but are payments for AI intelligences to cover costs for computing power, data, API calls, and other services. The transaction frequency and amounts are contrary to the design logic of bank cards. In September 2025, Coinbase and Cloudflare co-founded the x402 Foundation to promote industry-wide standards, rather than building closed, proprietary products.

In summary, by early 2026, five flagship AI commercial payment projects have been implemented: three using bank card settlement, and two using stablecoins, with application scenarios clearly divided between personal consumption and machine transactions.

Industry future outlook

In the short term, the industry landscape in 2026 is likely to remain as it is: bank cards dominate personal retail payments, stablecoins focus on machine-to-machine transactions, and both coexist and develop. But by 2030, this situation may be disrupted, as both camps are actively competing for the integration of these two scenarios.

The ultimate determinant will be whether AI-driven commercial transactions lean more toward traditional retail forms or evolve into vast networks of small-value machine transactions. If the former, traditional card organizations will continue to hold the dominant position; if the latter, stablecoin channels will capture a large share of new transaction flows.

Visa and Mastercard have made the most prudent choice: deploying both routes, so they can earn fees regardless of where future traffic flows. The real risk lies in companies that only bet on a single payment channel. Both card organizations have already mitigated this risk, which also reflects their judgment of the industry’s future.