OpenRouter has raised $113 million in a Series B round led by CapitalG, Alphabet’s independent growth fund, as the AI model routing platform’s weekly token volume hit 25 trillion. The round includes participation from Nvidia’s NVentures, ServiceNow Ventures, MongoDB Ventures, Snowflake Ventures, and Databricks Ventures, with existing investors Andreessen Horowitz and Menlo Ventures also returning, according to SiliconANGLE.
The company provides a unified API layer across 400+ AI models from both closed-source providers and open-source ecosystems. Developers and enterprises can route inference requests to different models based on task type, cost constraints, and latency requirements through a single interface. That routing layer also handles billing, usage tracking, access permissions, and failover across providers.
Volume and Adoption
OpenRouter’s weekly token volume has grown fivefold in six months, from 5 trillion tokens to 25 trillion, according to the company’s own usage data. The platform serves more than 8 million global users, including AI-native startups and large enterprises. Those rankings have become a de facto industry signal for real-world model adoption, as The New York Times noted in its coverage of the raise.
“Running inference at scale is fundamentally a multi-model problem. The era of picking a single model is over,” OpenRouter co-founder and CEO Alex Atallah told SiliconANGLE. “Success now depends on continuously routing across a changing market.”
The Enterprise Cost Problem
The raise lands as enterprise token consumption continues to climb. A Deloitte study cited by SiliconANGLE found that 67% of enterprise businesses already process nearly 1 billion tokens per month. Pricing varies dramatically across models: OpenAI’s GPT-5.5 charges $5 per million input tokens and $30 per million output tokens, while open-source alternatives like Qwen3.7 Max run at $2.50/$7.50 respectively.
That pricing spread creates an operational problem. Companies running autonomous agents, multimodal pipelines, or high-volume inference need to route different task types to appropriately sized models. Summarization tasks might go to a lightweight text model while complex reasoning gets routed to a frontier model. Building that routing logic in-house is expensive; buying it as a service is what OpenRouter sells.
The Strategic Investor Signal
The investor roster reads as a who’s who of enterprise data infrastructure. ServiceNow, MongoDB, Snowflake, and Databricks all participated through their venture arms, as did Nvidia through NVentures. For agent builders evaluating inference infrastructure, the signal is clear: the companies whose platforms generate and consume the most AI workloads are betting that model routing becomes a standard infrastructure layer, not a feature any single vendor owns.
OpenRouter said it will use the funding to expand routing, governance, and optimization capabilities as enterprises scale AI operations, according to Investing.com.