AI agents don't search the web like humans, and Exa thinks they deserve better than a Google wrapper. The startup just raised $250 million in Series C funding led by a16z to prove it.
The round gives Exa a post-money valuation of $2.2 billion — triple the $700 million it was worth last September.
Exa plans to use the money to expand its proprietary index, train next-generation retrieval models, and scale its backend to handle hundreds of thousands of concurrent requests per second.
The company's moat, as it tells it, is a 'full-stack self-built' approach: its own crawler, a native index of more than 500 billion URLs, in-house embedding models, and a third-generation vector database. No 'Google wrapper' architecture.
That purpose-built stack, Exa argues, is what AI search infrastructure should look like when it's designed from the ground up for AI agents.