Exa raises $250M Series C to build the search engine for AIs
AI agents will search the web more than humans this year.
Exa just raised $250M at a $2.2B valuation led by a16z to power all agents with the highest quality web search.
Exa already powers search for Cursor, Cognition, HubSpot, OpenRouter, Monday.com and over 400,000 developers. These are the early days - in the next few years the number of searches from LLMs will be 1000x more than Google searches today. This is an opportunity to redesign how the world gets information.
We started Exa five years ago to build perfect search over all the world’s information, far beyond traditional search engines in quality and comprehensiveness.
In early 2023, we saw that AI products and AI agents would most benefit from a perfect search API, so Exa announced the first web search API for AI. Since then, the number of companies using Exa has grown to over 5,000.
Customers choose Exa because we’re the highest quality search API at every latency and price point.
Some examples:
- GTM agents need extremely comprehensive search over people and companies, so we built comprehensive indexes over these verticals and can run for minutes to gather all the data agents need.
- Coding agents need complex searches over public code / technical documentation, so we finetuned special embedding models for code search.
- Everyone wants their chat agents to be faster and cheaper, so we built the fastest search API in the world (sub200ms) and fast text extraction models which reduce LLM token counts by over 20x.
This is only possible because we spent years building a web-scale search engine from scratch. Our crawlers track over 500 billion urls, our research teams train special embedding models on a GPU cluster we assembled, and we’ve built new vector databases for the extremely high QPS that agents need. There’s a reason there are more space programs than independent search engines.
Most other search providers actually wrap other search engines and therefore cannot compete on quality/latency/cost. As we scale up infra and model training in the coming months, the gap between Exa and wrappers will become clearer. For example, six months ago we were worse than Google at code search, and now we’re used by nearly every coding agent.
With this funding, we’re going to train our next-gen class of models and scale our infrastructure to handle 100s of thousands of searches per second. We’re assembling the best people around the world to do this. Recent hires include the head of retrieval infra from Meta, head of search backend at Yandex, and a research team out of Google.
We're also scaling our go-to-market org. Marcus Holm, President of LaunchDarkly, is joining as Exa’s CRO to lead our global GTM organization. We're hiring across sales, solutions, and marketing worldwide.
As trillions of agents come online over the coming years, search needs will grow thousands of times beyond the total search volume of Google. And as agents make increasingly important business decisions, their requirements for comprehensiveness, freshness, and precision will far exceed what humans require. In short, agents will need perfect search over all the world’s information at an unprecedented scale.
Building a perfect search engine will be extremely hard. It's also necessary. Information is critical civilizational infrastructure for our new AI reality. Politics is fragmenting, wars are raging, and AI is accelerating -- we need tools we can trust that deeply inform us what's going on. If we can build perfect search so that every AI has the highest quality information, then every human will too.
