Stripe preyed on OpenRouter. Now he doesn't have enough money, he also wants tokens.
On November 10, Stripe officially completed the acquisition of OpenRouter. The purchase agreement was announced on August 19, but the deal has now closed.
For those in the tank: OpenRouter is one of the largest aggregators of AI models. One API gives access to more than 400 models from 80+ vendors, including OpenAI, Anthropic, Google, and others.
According to OpenRouter itself, the platform is used by more than 10 million developers and companies, and more than 10 trillion tokens pass through it every day.
Why did Stripe need OpenRouter?
In my opinion, Stripe is not buying technology but so much buying a place in the future AI economy.
First of all, money. Stripe already knows how to accept payments, issue invoices and calculate the cost of using the services. Now we will add to this the routing of requests between neural networks. The customer can be offered everything at once: access to models, cost optimization and billing.
Second, controlling AI costs. When a company uses dozens of models, choosing between them becomes an economic challenge. Where is it cheaper? Where is faster? Which model will better handle a specific request? Stripe wants to be the infrastructure that helps make these decisions. But they still have zero expertise.
Third, AI agents. If autonomous agents make billions of transactions, someone will have to account for every token issued, process payments and manage budgets. Clearly, Streep is looking to fill that void.
And finally, market data. OpenRouter sees which models developers choose, how demand changes and how much inferences cost. For Stripe, this is potentially valuable market intelligence.
And now the most elegant part.
OpenRouter solemnly promises to remain neutral: models and providers will continue to compete on quality, reliability and price. The name, product and existing integrations will also be preserved.
That is, everything will remain independent and neutral.