Exa, a San Franscisco-based startup focused on building a search engine for AI agents, has secured $250m in Series C funding at a valuation of $2.2bn.
The round was led by venture capital firm Andreessen Horowitz (a16z).
Go deeper with GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
Exa intends to use the new capital to further develop web search infrastructure targeted at AI applications.
Its search engine currently serves over 5,000 companies and more than 400,000 developers, including firms such as HubSpot, OpenRouter, Monday.com, Cursor, and Cognition.
Founded five years ago, Exa created a search platform intended to retrieve comprehensive information from the web for various clients.
In early 2023, Exa introduced its web search API for AI-focused products, aiming to deliver search capabilities to support the growing demand from AI agents.
The company reports that demand from AI products for search is expected to surpass that of traditional engines like Google by a significant margin in the coming years.
Exa’s search system was built in-house, including custom crawlers that monitor over 500 billion URLs and proprietary embedding models trained on an internal GPU cluster.
The system also uses newly developed vector databases engineered to handle high query per second (QPS) workloads required by AI-driven agents.
According to Exa, many other providers rely on wrapping existing search engines and lack the same control over quality, latency, and costs.
The company plans to use the Series C proceeds to train more advanced models and expand infrastructure capable of handling hundreds of thousands of searches per second.
Exa has also expanded its global team, making hires from major technology companies, including Meta, Yandex, and Google.
In addition, Marcus Holm, president of LaunchDarkly, has joined Exa as its chief revenue officer to oversee go-to-market efforts.
Exa says it has developed specialised search solutions for different applications, including comprehensive people and company indexes for go-to-market agents and improved code search for coding agents. The company has also created rapid search APIs designed to reduce operational costs and latency for chat agents.
