Weaviate is an open-source, AI-native vector database founded in 2019 by Bob van Luijt and Etienne Dilocker in Amsterdam, the Netherlands, operating as Weaviate B.V. Originally conceived as a semantic search engine, it has matured into one of the most widely deployed vector databases for retrieval-augmented generation, semantic search, and multimodal retrieval, with a permissive BSD-licensed core and a global open-source community.
Weaviate stores and queries high-dimensional embeddings alongside structured metadata. Its built-in vectorization modules can call out to OpenAI, Cohere, Hugging Face, Anthropic, and other model providers to embed data at ingest time, removing the need for an external embedding pipeline. The engine supports dense vector search, BM25 keyword search, and hybrid queries that blend both, with a GraphQL-style API and a newer REST and gRPC interface.
The company raised a $50M Series C in October 2025 led by Battery Ventures with participation from Zetta Venture Partners and existing investors including Index Ventures and NEA, reportedly at a $200M post-money valuation. This builds on a 2023 Series B of the same size led by Index Ventures, bringing total funding above $100M.
Weaviate is offered as self-hosted open source, as a fully managed Serverless Cloud, as a dedicated Enterprise Cloud across AWS, GCP, and Azure, and as bring-your-own-cloud and on-prem deployments. The product roadmap has leaned hard into multi-tenancy, ANN compression to reduce memory footprint, agentic retrieval primitives, and tighter integrations with LangChain, LlamaIndex, and Haystack.
Weaviate competes with Pinecone, Qdrant, Milvus, and pgvector-on-Postgres. It differentiates on its AI-native module system, hybrid search quality, and a deliberately open-source-first posture — important for regulated enterprises and developers who want to avoid lock-in to a closed managed service.