AWS has launched Amazon S3 Vectors in preview, the first cloud object storage service with native vector support at scale. Designed for generative AI workloads, S3 Vectors enables affordable storage, subsecond query performance, and up to 90% cost reduction for uploading, storing, and querying vector embeddings.
Vectors, numerical representations of unstructured data generated by embedding models, are key to powering semantic and similarity search.
With this launch, AWS brings a durable, purpose-built solution that allows developers to manage massive AI-ready vector datasets directly within Amazon S3, significantly simplifying architecture for applications that rely on embedding-based search and retrieval.