Technology

Amazon S3 Vectors

A native vector storage and search capability built into S3, enabling storage and querying of embeddings directly in S3 without a separate vector database. Supports up to 20 trillion vectors per bucket and 2 billion per index as of late 2025.

7 connections 3 resources

Summary

What it is

A native vector storage and search capability built into S3, enabling storage and querying of embeddings directly in S3 without a separate vector database. Supports up to 20 trillion vectors per bucket and 2 billion per index as of late 2025.

Where it fits

S3 Vectors integrates vector search into the S3 storage layer, reducing the need for standalone vector databases for large-scale, cost-optimized similarity search. At 20 trillion vectors per bucket, query federation across multiple smaller indexes is no longer necessary for most use cases.

Misconceptions / Traps
  • Not a full vector database replacement for all use cases. Optimized for scale and cost, not for ultra-low-latency real-time search scenarios where in-memory indexes are needed.
  • Query performance characteristics differ from dedicated vector databases. Evaluate latency requirements before choosing S3 Vectors over purpose-built solutions.
  • The 2-billion-per-index limit means very large deployments may still need multiple indexes within the same bucket, but bucket-level federation is eliminated.
Key Connections
  • enables Semantic Search — native vector search capability in S3
  • enables Hybrid S3 + Vector Index — embeddings and raw data co-located in S3
  • scoped_to Vector Indexing on Object Storage — S3-native implementation
  • constrained_by Vendor Lock-In — AWS-specific feature

Definition

What it is

Native vector storage and similarity search capability built into Amazon S3, allowing embeddings to be stored and queried directly within object storage without a separate vector database. Supports up to 20 trillion vectors per bucket and 2 billion per index as of late 2025, effectively removing the need for query federation across multiple smaller vector indexes.

Why it exists

Traditional vector search requires a dedicated vector database (Pinecone, Milvus). S3 Vectors eliminates this separate layer by integrating vector storage and search directly into S3, reducing infrastructure complexity.

Primary use cases

Serverless semantic search over S3-stored data, RAG without a separate vector DB, vector similarity queries at S3 scale, trillion-scale embedding stores without index federation.

Connections 7

Outbound 6
Inbound 1

Resources 3