Stripe Buys OpenRouter to Help AI Companies Manage Token Spending

Stripe agreed last week to acquire OpenRouter, the AI model gateway that routes requests across more than 400 models from over 80 providers. Neither company disclosed terms. The New York Times reported a $7.5 billion price, while Axios and Reuters put the value above $8 billion.

OpenRouter doesn’t train models. It gives developers one interface to every major model and routes each request based on task complexity, price, speed and reliability. In its own announcement, the company said it processes more than 10 trillion tokens a day for over 10 million developers and companies, and that inference volume has grown at least tenfold every year since it launched in early 2023. Stripe says customers include NVIDIA, Zoom and Lovable.

Why a payments company wants a model router

Stripe already processes payments for most AI companies and has spent the past year building tools such as Token Billing to help them manage inference costs. Owning the routing layer lets it work on both sides of an AI company’s margin: the revenue it collects and the tokens it spends.

“Tokens are the central currency for companies building with AI, and it’s clear that the real-world economic potential will depend on making good use of scarce compute resources,” said Patrick Collison, Stripe’s cofounder and CEO. “Stripe is building the economic infrastructure for AI, and together with OpenRouter we’ll help businesses maximize profitability by routing their requests intelligently and spending their tokens efficiently.”

Neutrality is the promise

For developers, the big question is whether OpenRouter stays neutral under a new owner. The company says it will operate as before, with the same name, product and roadmap, and that nothing about existing integrations changes. “Routing decisions will remain driven by one thing: what’s best for you, the user,” it wrote.

Cofounder and CEO Alex Atallah framed the deal around a multi-model future. “We believe intelligence will be multi-model: no single model will be optimal for every task, and developers need a neutral layer to orchestrate and manage them all,” he said. For companies whose inference bills now rank among their biggest expenses, a well-funded neutral router could make shopping across models easier.