Visa and Stripe-Backed Tempo Unveil New Plans for AI Payments Infrastructure

A new wave of financial infrastructure is emerging as artificial intelligence begins to handle not just tasks—but transactions. Payments giant Visa and the Stripe-backed blockchain project Tempo have both introduced new tools designed to let AI agents pay for services autonomously online.

At the center of Visa’s push is a newly introduced experimental product from its crypto division, known as Visa CLI. Developed by Cuy Sheffield and his team, the tool gives AI agents the ability to execute payments in real time while performing tasks such as writing code or interacting with digital services.

Unlike traditional payment integrations, Visa CLI removes the need for API keys—long considered a weak point in automation due to potential security leaks. Instead, it enables secure, programmatic card payments directly within an agent’s workflow, allowing AI systems to transact without exposing sensitive credentials.

At the same time, Tempo has launched its blockchain on mainnet, introducing a payments protocol tailored specifically for AI-driven use cases. Built with backing from Stripe, the network focuses on handling high-volume stablecoin transactions—currently one of the most common ways AI agents interact with financial systems.

Tempo’s broader vision goes beyond infrastructure. The project also unveiled the Machine Payments Protocol, an open standard co-developed with Stripe. This framework is designed to allow AI agents and digital services to coordinate payments seamlessly, regardless of the payment method being used.

The protocol is intentionally flexible, supporting multiple rails including cards, digital wallets, and stablecoins. Visa has already extended compatibility through its card network, while Stripe is enabling broader payment method support across its ecosystem. Meanwhile, Lightspark has integrated the protocol with the Bitcoin Lightning Network, expanding its reach into crypto-native payment channels.

This latest development builds on a growing trend toward agentic payments. Earlier efforts, such as the x402 standard introduced by Coinbase and later integrated into tools by Sam Altman’s World project, have already begun laying the groundwork for AI-native financial interactions.

As AI agents become more capable—handling everything from coding to coordinating services—the need for seamless, automated payment systems is becoming unavoidable. Industry players are increasingly betting that machine-to-machine transactions could soon rival, or even surpass, human-initiated payments on the internet.