menu
close

AWS Launches S3 Vectors, Slashing AI Storage Costs by 90%

Amazon has introduced S3 Vectors, the first cloud object storage with native vector support for AI workloads. This groundbreaking solution reduces the cost of storing and querying vector data by up to 90% compared to conventional methods, while providing sub-second query performance. S3 Vectors seamlessly integrates with Amazon Bedrock Knowledge Bases and other AWS services, making it cost-effective to leverage large vector datasets for AI applications and semantic search.
AWS Launches S3 Vectors, Slashing AI Storage Costs by 90%

Amazon Web Services (AWS) has unveiled Amazon S3 Vectors, a purpose-built durable vector storage solution that promises to transform how organizations store and utilize AI data at scale.

Announced on July 15, 2025, at the AWS Summit in New York, S3 Vectors is the first cloud object store with native support for storing and querying vector embeddings. The service can reduce the total cost of uploading, storing, and querying vectors by up to 90% compared to traditional vector databases, while maintaining sub-second query performance.

Vector embeddings, which are numerical representations of unstructured data created from embedding models, have become essential for modern AI applications. They enable semantic search capabilities and provide context for large language models. However, conventional vector storage solutions typically require dedicated compute resources that run continuously, driving up costs significantly.

"When we looked across customer workloads, we found that the vast majority of vector indexes did not need provisioned compute, RAM or SSD 100 percent of the time," explained AWS in their announcement. For example, a conventional vector database with ten million vectors can cost over $300 monthly on a dedicated instance, whereas the same dataset in S3 Vectors would cost approximately $30 per month with 250,000 queries.

S3 Vectors introduces a new bucket type with dedicated APIs for vector operations, allowing users to store and query vector data without provisioning infrastructure. Each vector bucket can contain up to 10,000 vector indexes, with each index capable of holding tens of millions of vectors. The service automatically optimizes vector data for the best possible price-performance ratio, even as datasets scale and evolve.

The solution integrates natively with Amazon Bedrock Knowledge Bases, Amazon SageMaker, and Amazon OpenSearch Service, making it particularly valuable for retrieval-augmented generation (RAG) applications. Organizations can implement a tiered strategy, storing large vector datasets in S3 for cost efficiency while moving frequently accessed vectors to OpenSearch for higher performance when needed.

S3 Vectors is currently available in preview, with AWS inviting customers to try it through the Amazon S3 console.

Source:

Latest News