Stripe aims to power the next generation of AI-driven payments
Image Source: depositphotos.com
Stripe has not only become a leader in modern online payment systems in just a few years, but has also set itself an even more ambitious goal: moving from e-commerce to AI commerce.
Stripe wants to become the leading payment infrastructure for AI agents, and it has made that ambition clear with the acquisition of OpenRouter, a platform that routes requests from users and developers to more than 400 AI models from over 80 providers.
The deal reportedly cost around $7.5 billion, but Stripe did not stop there. Just a few weeks later, it announced that Link would allow Muse, Meta’s new personal AI agent, to make purchases across the web. In the United States, Muse can use the payment method stored by a user in Link to shop at more than one million businesses that accept it.
This is where so-called “agentic commerce” comes in. Agents will start searching for products, comparing prices, and making purchases on behalf of users. In this environment, the infrastructure supporting these activities must protect user privacy while managing identities, payments, and machine-to-machine transactions.
The company has also increased its exposure to stablecoins, a technology that could prove useful in its effort to become the financial infrastructure behind the AI agent economy, with stablecoins potentially gaining more ground on the crypto heatmap as the market evolves.
Such autonomy, however, also brings risks. When should an agent stop? What is its spending limit? Who authorizes the transaction, and above all, who manages access to the user’s financial credentials?
Stripe is positioning itself to answer these questions by building a financial infrastructure capable of setting boundaries for agents without compromising their functionality. It has already developed a Link wallet for agents, which enables an agent to receive a single-use card or a Shared Payment Token without accessing the user’s actual card details. New protocols are being developed to support direct micropayments and stablecoin transactions between agents.
The company is also working with OpenAI on the Agentic Commerce Protocol and has developed its own Agentic Commerce Suite, a system that allows merchants to make their products available to different AI agents without having to build a separate integration for each one.
In this context, OpenRouter is strategically important because it allows applications to access multiple models through a single interface and select the most suitable option based on price, performance, and availability.
Agentic commerce remains a major challenge in terms of costs and business risks. It remains unclear how practical the model will be at scale, what its actual operating costs will look like, and whether it can be made sufficiently secure for consumers. Fraud, errors, refunds, and unauthorized transactions are all issues that could slow its development or even make the model commercially unviable, a risk that would weigh directly on any future Stripe IPO valuation.