Exa Labs Inc.
Exa is an AI research lab building the search engine for artificial intelligence. Their mission is to organize all the world's knowledge and achieve 'perfect search' where information is fully organized and accessible for AI applications, enabling AI agents to retrieve accurate, relevant web data without hallucinations.
Founding Story
Exa was founded in 2021 by Will Bryk and Jeffrey Wang, who were roommates at Harvard University. They built their first search engine together five years before founding Exa while studying Computer Science at Harvard. The company started in the summer 2021 Y Combinator cohort, initially intending to use AI to build better search for humans. After launching their first search engine product in November 2022 (just before ChatGPT's release), they realized that AIs would eventually search the web more than humans. This insight led them to pivot and focus on becoming the primary search infrastructure for AI. They built a proprietary GPU cluster and indexing system from scratch rather than relying on existing search providers like Google or Bing. The company was formerly known as Metaphor before rebranding to Exa in early 2024.
Discussions
No discussions yet
Be the first to start a discussion about Exa Labs Inc.
Leadership
Founders
Will Bryk
CEO and Co-Founder. Previously one of the first engineers at Cresta where he built real-time AI products. Studied Computer Science and Physics at Harvard University. Age 27. Expert in embedding models, researched human/AI interaction, and led robotics club at Harvard.
Jeffrey Wang
Co-Founder. Previously worked on data and web infrastructure at Plaid. Studied Computer Science and Philosophy at Harvard University. Age 26. Expert in social analysis. Built his first search engine with Will Bryk while roommates at Harvard.
Executive Team
Will Bryk
CEO and Co-Founder
Former engineer at Cresta, studied CS and physics at Harvard
Jeffrey Wang
Co-Founder
Former data and web infrastructure engineer at Plaid, studied CS and Philosophy at Harvard
Business Model
Revenue Model
API usage-based pricing model with pay-as-you-go and custom enterprise plans. Charges based on number of API requests, with different pricing for different endpoint types (Search, Contents, Answer, Research, Crawl). Revenue generated from API calls by developers, AI companies, and enterprises.
Pricing Tiers
For individuals and small teams. Includes $10 in free credits, no credit card required. Search (fast/auto/neural 1-25 results): $5/1k requests; (26-100 results): $25/1k requests; Deep search: $15/1k requests; Contents (text/highlights/summary): $1/1k pages; Answer: $5/1k answers; Research: variable pricing based on operations.
For high volume, custom datasets, and enterprise security. Includes volume discounts, standard billing, up to 1,000 results per search, tailored moderation, 1:1 onboarding, SLAs, MSAs, comprehensive security with zero data retention, custom rate limits (QPS), and early access to new features.
Target Markets
- AI startups and developers
- Enterprise AI applications
- Consulting firms (Big 3 and others)
- Private equity firms
- Investment banking and financial services
- Marketing agencies
- AI coding assistants retrieving technical documentation and code examples
- AI chatbots accessing real-time information to answer customer questions
- Market research and competitive analysis for consulting firms
- Training data curation for AI models
- Financial analysis and historical market data retrieval
- Recruiting and candidate profile searches
- Cursor
- Notion AI
- Databricks
- AWS
History & Milestones
Released exa-d: Data Framework to Process the Web
Revealed Exacluster - one of the industry's first clusters based on 144 Nvidia H200 GPUs
Raised $85M Series B at $700M valuation led by Benchmark
Launched Exa 2.0
Launched Exa 2.1
