Exa Bags $85 Million Series B to Build the Search Engine for AIs

Exa, a Search API Platform for AI Systems, led by Will Bryk, Jeff Wang, Thais Branco, and others, has raised $85 million in Series B funding led by Benchmark with participation from Lightspeed, YCombinator, and NVentures, NVIDIA's venture capital arm. Peter Fenton from Benchmark will join the board.

Exa already powers web search for thousands of companies, from AI startups like Cursor to leading private equity and consulting firms. Built from scratch, its search engine delivers state-of-the-art speed and quality, optimized for AI applications.

A search engine for AI is built specifically to meet the unique needs of artificial intelligence systems, which, unlike humans, require fast, comprehensive, and high-quality access to information rather than SEO-driven results or ads. Exa is designed for this purpose, delivering optimized ranking for reliable knowledge, full-page content instead of just links and titles, and sub-450ms latency for speed-critical applications. For deeper research, its high-compute product, Websets, enables large-scale, comprehensive searches. It is also highly customizable to different AI use cases and ensures Zero Data Retention for enterprise security. This makes it a foundational tool for AI applications ranging from startups to global financial and consulting firms.

Exa is entering its second stage with the goal of achieving perfect search, organizing the world’s information far beyond existing systems. The next steps include massively scaling indexing and processing, expanding GPU capacity fivefold to accelerate research, and growing the team across engineering, GTM, and operations. Exa offers two products, the API and Websets, both available to try at exa.ai.