Voyage AI
Category: AI Infrastructure
Voyage AI builds state-of-the-art embedding and reranking models that power retrieval-augmented generation (RAG) and enterprise LLM applications. Voyage AI was founded in 2023. The company is led by Tengyu Ma. Based in Palo Alto, California, USA. Team size: 11-50. Total funding raised: $28M. Latest round: Acquired. Key investors include CRV, Conviction, Wing VC, Snowflake, Databricks, Pear VC, Tectonic Ventures, Mayfield Fund, Fusion Fund.
- Founded
- 2023
- Headquarters
- Palo Alto, California, USA
- Team size
- 11-50
- Total funding
- $28M
Value proposition
Delivers highly accurate, scalable embedding and reranking models that make enterprise information retrieval faster and more reliable, reducing hallucination in LLM applications.
Products and solutions
Voyage embedding models (voyage-3, voyage-3-large, voyage-code, voyage-law, voyage-finance domain-specific embeddings), Voyage rerankers (rerank-2), API and SDK for retrieval pipelines, now integrated into MongoDB Atlas Vector Search.
Unique value
Domain-specialized, state-of-the-art embedding and reranking models built by Stanford/MIT/UC Berkeley academics plus industry engineers from Google, Meta, Uber, achieving top MTEB benchmark results and outperforming OpenAI/Cohere on several retrieval benchmarks.
Target customer
Enterprises building RAG pipelines and production AI applications; developers integrating semantic search/retrieval into LLM products.
Industries served
Enterprise software, legal, finance, code/developer tools, general RAG/LLM applications across verticals
Technology advantage
Academic research pedigree (Stanford NLP/ML lab lineage) combined with production-grade model training; strong benchmark performance (MTEB leaderboard); domain-specific fine-tuned embedding models (legal, code, finance) that generalist providers lack.
How they differentiate
Focuses specifically on best-in-class embedding/reranking quality with domain-specialized models (legal, code, finance) rather than being a generalist LLM provider; strong academic pedigree drives benchmark-leading MTEB/NDCG performance.
Main competitors
Cohere (embed-v4), OpenAI (text-embedding-3-large), Jina AI (jina-embeddings)
Key partnerships
MongoDB (acquirer, deep product integration into Atlas Vector Search), Snowflake and Databricks (investors and likely data-platform integration partners)
Notable customers
MongoDB Atlas Vector Search users (post-acquisition integration), enterprise RAG customers (specific names not publicly disclosed)
Major milestones
2023: Founded by Tengyu Ma and academic/industry team, Oct 2024: Raised $28M total (Series A $20M led by CRV), Feb 24, 2025: Acquired by MongoDB (NASDAQ: MDB) for approximately $220M in a cash-and-stock deal.
Market positioning
Premium, quality-leading embedding/reranking model provider for enterprise RAG, now embedded as MongoDB's native AI retrieval layer post-acquisition.
Geographic focus
Primarily US-based enterprise customers, competing globally against OpenAI, Cohere, and Jina AI in the embedding model market.
About Tengyu Ma
Assistant Professor at Stanford University (Computer Science); PhD Princeton. Co-founded Voyage AI in 2023; after MongoDB's 2025 acquisition, now serves as Chief AI Scientist at MongoDB while remaining a Stanford professor.
Latest news about Voyage AI
More AI Infrastructure companies
Official website: https://www.voyageai.com