Corporate expense management company Ramp is entering the AI infrastructure race. On Wednesday evening, the company launched Router, an API-based service that lets businesses route their AI queries across multiple large language models from providers including OpenAI, Anthropic, DeepSeek, Moonshot, Minimax, Nvidia, xAI, and Z.ai. The service is free for the remainder of 2026 and comes with a $26 credit for new users, though Ramp has not disclosed pricing for next year.
Router is available only in the United States. Ramp says it has been using the internal version of the tool for its own AI needs for the past three years, and is now offering it to customers as a standalone product.
Also read: Pew Study: 35% of Post-ChatGPT Webpages Show Signs of AI Authorship
What Ramp Router does differently
At its core, Router functions similarly to OpenRouter, a popular model gateway that aggregates dozens of AI models under a single API. But Ramp’s version is more limited in model selection, and instead focuses on giving users fine-grained control over how requests are routed.
Router offers several “strategies” that let customers set preferences. For example, one strategy routes queries to providers’ flex usage tiers, which can lower costs in exchange for variable latency. Another lets users specify up to three benchmarks, and Router will choose the model that best fits those criteria. Users can also direct only difficult problems to premium models, or test new models without switching their entire stack.
Also read: Inertia Enterprises slashes fusion fuel production time from days to hours
The service includes a dashboard that shows token spend, cost, latency, fallback attempts, and other operational details. That data layer aligns with Ramp’s existing AI token usage monitoring and spend management tools, which are already part of its expense platform.
Why Ramp is moving into AI infrastructure
Ramp’s move comes just weeks after Stripe reportedly agreed to acquire OpenRouter for more than $7 billion, a deal that signaled the growing value of AI gateway infrastructure. By offering its own router, Ramp is positioning itself to capture a slice of the booming AI inference market, which is projected to grow rapidly as enterprises deploy more generative AI applications.
For Ramp, the opportunity is two-pronged. The company can generate new revenue from API usage, and it can deepen relationships with existing clients who already use its expense management tools. Router also gives Ramp a new entry point to sell its core products to developers and AI teams who may not have considered the company before.
The company raised $750 million at a $44 billion valuation in June, giving it ample resources to experiment with new product lines. If Router gains traction, it could also help Ramp build long-term partnerships with AI labs and inference providers, similar to how OpenRouter has become a testing ground for new models.
Data retention and privacy considerations
Router has an opt-out data retention policy. By default, Ramp will record model inputs, outputs, and tool calls for one year. The company says it will remove personally identifiable information before using that content to improve the product, but users who want stricter controls can opt out of data retention entirely.
That policy may raise questions for enterprises with strict data governance requirements, especially those handling sensitive financial or customer data. Ramp’s existing clients in the finance and fintech sectors will likely scrutinize how Router handles data before integrating it into their workflows.
What to watch next
Ramp has not said when Router will expand beyond the United States or what its pricing will look like in 2027. The company also has not indicated whether it plans to add more models to the roster, which currently trails OpenRouter’s catalog significantly.
For now, Router is a free tool that lets developers experiment with model routing without committing to a paid plan. That could make it an attractive option for startups and mid-sized companies looking to manage AI costs more effectively. Whether Ramp can turn that interest into a lasting business remains an open question, but the company’s existing enterprise relationships give it a solid foundation to build on.
This article is for informational purposes only and does not constitute financial advice. The AI model routing market is volatile and evolving rapidly; any forward-looking statements about Ramp’s strategy or pricing are speculative and subject to change.
Be the first to comment