OpenRouter Completes $113 Million Series B Financing, Led by CapitalG
OpenRouter announced the completion of $113 million in Series B financing, led by CapitalG, a subsidiary of Google's parent company Alphabet. NVentures, ServiceNow Ventures, MongoDB Ventures, Snowflake Ventures, and Databricks Ventures participated, with follow-on investments from existing shareholders such as a16z and Menlo Ventures.
OpenRouter is an AI model routing platform that allows enterprises to access over 400 models, including those from Anthropic, Google, OpenAI, xAI, and DeepSeek, through a single API.
The platform's weekly processing volume has reached 250 trillion tokens (approximately 1 quadrillion per month), a fivefold increase from six months ago, currently serving over 8 million users globally.
Source: Public Information
ABAB AI Insight
OpenRouter previously focused on model aggregation routing. This round, led by CapitalG, continues its path towards deepening enterprise-level AI infrastructure, upgrading from simple API calls to providing unified routing, access control, spending audits, and governance capabilities to address the complexity of multi-model scheduling for enterprises.
In terms of capital strategy, OpenRouter will focus the new funding on routing optimization and governance system construction, aiming to create data barriers through real call traffic accumulation. The goal is to become an intermediary control layer for enterprise AI procurement, helping clients dynamically balance costs, latency, performance, and compliance while securing traffic distribution rights in a multi-model era.
Similar to Databricks' positioning in the data field and Snowflake's enterprise data platform strategy, OpenRouter is currently in an expansion phase from the single-model era to the multi-model intelligent scheduling era.
Essentially, this represents capital concentration: as model vendors engage in fierce competition, enterprises increasingly rely on an intermediary routing layer to manage traffic and costs uniformly. The mechanism is that real token call data becomes a scarce asset, prompting capital to concentrate on platforms that possess scheduling rights and governance capabilities, forming new control points in the AI supply chain.
ABAB News · Cognitive Law
Model vendors compete for capabilities, while routing platforms earn traffic and governance fees.
The more models enterprises use, the more they rely on a unified scheduling intermediary layer.
Whoever controls real calls stands at the top of the AI procurement chain.