AI can now pay on its own as Visa, Stripe roll out new rails

Artificial intelligence is steadily moving beyond automation into economic activity, with new tools now allowing software agents to pay for services directly.

On Wednesday, Visa’s crypto unit introduced an experimental product designed to let AI systems execute payments while coding.

The same day also saw the launch of the Tempo blockchain, backed by Stripe, alongside a payments protocol aimed at enabling machine-to-machine transactions.

Together, the developments reflect growing efforts to standardise how AI agents interact with financial infrastructure as stablecoin usage and agent-driven workflows expand.

Visa crypto tool

Visa’s crypto division unveiled Visa CLI, a command-line interface tool that enables artificial intelligence agents to initiate payments during development workflows.

The announcement was shared on Wednesday by Cuy Sheffield via X, positioning the tool as the first experimental release from Visa Crypto Labs.

The product is designed to allow developers to instruct AI systems to complete payments directly while writing code.

According to its website, the interface enables secure transactions without relying on traditional API keys.

These keys often carry sensitive credentials that could be exposed if handled by AI systems, raising security concerns.

By removing the need for API keys, the tool aims to reduce risks linked to data leakage while enabling seamless integration of payments into AI-driven workflows.

The system supports programmatic card payments, allowing agents to execute transactions based on predefined instructions.

Agent payments race

Visa’s launch comes amid a broader push to define standards for agentic payments.

As AI systems become more capable of executing tasks independently, the need for built-in financial functionality has gained attention across the crypto and fintech sectors.

Coinbase introduced its x402 standard in May to support stablecoin-based payments between AI agents.

That framework was later integrated into a developer toolkit released on Tuesday by Sam Altman’s World project, signalling increasing adoption of programmable payment rails for autonomous systems.

These developments highlight how stablecoins are emerging as a preferred medium for AI-driven transactions, given their speed and compatibility with programmable environments.

Tempo blockchain launch

At the same time, the Tempo blockchain, backed by Stripe, went live on mainnet with a focus on payments infrastructure tailored for AI systems.

The network is built to handle high-throughput stablecoin transactions, reflecting demand for scalable payment solutions.

Tempo stated on X that its blockchain is purpose-built for payments and designed to support the evolving needs of AI agents.

As these systems increasingly handle tasks such as writing code, coordinating services, and retrieving data, the ability to transact has become a necessary extension of their functionality.

Machine payments protocol

Alongside its launch, Tempo introduced the Machine Payments Protocol, developed in collaboration with Stripe.

The protocol provides a standardised framework for agents and services to coordinate payments programmatically.

The system is designed to remain rail-agnostic and adaptable, allowing compatibility across multiple payment methods.

Visa has extended support for the protocol through its card network, while Stripe enables integration with cards, wallets, and other payment options.

Lightspark has also added support for the protocol via the Lightning Network, enabling Bitcoin-based transactions.

This expands the range of payment rails available to AI systems, combining traditional card networks with blockchain-based infrastructure.

As these standards develop, the ability for AI agents to initiate and complete transactions is becoming a central focus across both crypto and traditional payments ecosystems.