Ramp enters AI inference market with Router launch

Corporate expense platform Ramp has unveiled Router, an API-based service allowing companies to switch between various large language models. The move mirrors industry trends toward centralized AI management, with Ramp leveraging its internal infrastructure to offer users a dashboard for monitoring token spend, latency, and model performance metrics.

Business Faces·August 20, 2026·reads: 512

Ramp enters AI inference market with Router launch

The service is currently limited to the United States and remains free for the remainder of 2026, though users must cover their own inference costs. Ramp is offering a $26 credit to encourage adoption. While the company has not disclosed pricing for 2027, the platform provides access to models from major providers including OpenAI, Anthropic, DeepSeek, Nvidia, and xAI.

Router functions through configurable strategies that allow companies to automate model selection based on specific benchmarks or task complexity. Users can direct difficult queries toward premium models while routing simpler requests elsewhere. The platform also includes a data retention policy that logs inputs and outputs for one year, though Ramp claims it strips personally identifiable information from this data before using it for product improvements.

For Ramp, which reached a $44 billion valuation following a $750 million funding round in June, the service serves as both a revenue stream and a client acquisition tool. By integrating AI routing into its established expense management ecosystem, the company aims to deepen its footprint in the enterprise AI space while building relationships with global model labs and inference providers.