Vector database
Also known as: vector store, vector index, vector search engine
A database optimised for storing and querying high-dimensional vectors (embeddings) using approximate-nearest-neighbour (ANN) search. The retrieval primitive of RAG-based agents. 2026 production options: Pinecone, Weaviate, Qdrant, Milvus, Chroma, plus pgvector (Postgres), OpenSearch (AWS), Vertex AI Search (Google), Azure AI Search.
Most 2026 enterprise agent deployments do not need a dedicated vector database. Postgres + pgvector handles up to ~10M vectors with reasonable latency on commodity hardware; Azure AI Search and OpenSearch ship vector primitives in the same managed product the deployment is already using for keyword search. Procurement teams that adopt Pinecone or Qdrant out of the gate often discover at scale that the operational simplicity isn't worth the additional vendor dependency. The right default in 2026 is to start with the in-house cloud's vector primitive and graduate to a specialist only when the workload's latency-at-scale numbers demand it.
Articles that analyse this term
Primary sources
- PostgreSQL. pgvector extension
- AWS. Bedrock Knowledge Bases — vector store options