Amazon S3 Vectors
Cost-optimized AI-ready storage with native support for storing and querying vectors at scale, reducing total costs by up to 90%
What is S3 Vectors?
Amazon S3 Vectors is the first cloud object store with native support to store and query vectors, delivering purpose-built, cost-optimized vector storage for AI agents, AI inference, and semantic search of your content stored in Amazon S3. By reducing the cost of uploading, storing, and querying vectors by up to 90%, S3 Vectors makes it cost-effective to create and use large vector datasets to improve the memory and context of AI agents as well as semantic search results of your S3 data. Designed to provide the same elasticity, scale, and durability as Amazon S3, S3 Vectors lets you store up to billions of vectors and search data with sub-second query performance. It's ideal for applications that need to build and maintain vector indexes at scale so you can organize and search through massive amounts of information.
Build faster with AI-ready storage
With S3 Vectors, you benefit from a dedicated set of APIs to store, access, and query vectors without provisioning any infrastructure. S3 Vectors is natively integrated with Amazon Bedrock Knowledge Bases, including within Amazon SageMaker Unified Studio, to reduce the cost of retrieval augmented generation (RAG). Through its integration with Amazon OpenSearch Service, you can adopt a tiered strategy to store large vector datasets in S3 for near real-time access while effortlessly activating the vector data with the highest performance requirements in OpenSearch.
Data Points
90%
lower costs for storing, uploading, and querying vectors2B
max vectors stored and queried per index100ms
lowest warm query latency performance10,000
indexes per bucket, up to 20 trillion vectorsBenefits
Reduce the cost of uploading, storing, and querying vectors by up to 90% while maintaining sub-second query performance. Transform your economics of storing millions to billions of vectors by moving away from costly storage options and paying only for what you use. Efficiently scale massive amounts of vectors without infrastructure management, organizing data using vector indexes that accommodate evolving workloads with zero provisioning. Designed for vector-driven AI use cases, S3 Vectors offers a practical balance of performance and efficiency.
Generate fine-grained vector embeddings to gain deeper understanding from unstructured data including images, videos, audio, and text. Elastically scale for vector search applications to improve granularity based on semantic similarity. Whether analyzing news content, indexing sports highlights, or working with medical images and genomic data, S3 Vectors supports high-volume workloads with consistent query performance and flexible scaling.
Use S3 Vectors for large, long-term vector data that doesn't require the high-throughput performance of in-memory vector databases. While Amazon OpenSearch Service delivers the high-QPS (query per second), low-latency vector search needed for real-time applications, S3 Vectors complements this by providing a cost-optimized data foundation with query performance optimized for long-term storage and infrequent access of data. You also benefit from a storage architecture with strong consistency guarantees, ensuring subsequent queries always include your most recently added data.
Leverage built-in connectivity with Amazon OpenSearch Service for vector search at optimized cost-performance and Amazon Bedrock Knowledge Bases for enhanced RAG applications at reduced costs. Access Amazon Bedrock within Amazon SageMaker Unified Studio to build inference-driven applications using existing project profiles, creating an integrated, scalable, and shareable AI development environment for enhanced team collaboration.
Use cases
Pinpoint search results based on semantic meaning and similarity
Perform semantic and similarity searches across large volumes of vector datasets. Media organizations can index millions of hours of video to instantly surface relevant scenes for highlight reels, while healthcare providers can store billions of vector embeddings representing medical images to identify similar cases and accelerate diagnosis. With S3 Vectors, you can unlock the semantic value of unstructured data at a lower cost without compromising scale.
Reduce RAG costs with Amazon Bedrock integration
Lower the cost of Retrieval Augmented Generation (RAG) by combining S3 Vectors with Amazon Bedrock Knowledge Bases. Turn your proprietary datasets into intelligent knowledge stores with contextual awareness by using your RAG applications. Quickly build and customize generative AI applications with access to scalable vector data in S3 Vectors as well as high-performing foundation models and advanced knowledge bases in Amazon Bedrock through its console, APIs, SDKs, or directly within Amazon SageMaker Unified Studio.
Build smarter AI agents with expanded and lasting memory
Make your AI agents more intelligent by retaining more context, reasoning with richer data, and building lasting memory from affordable, large-scale vector storage. Store every interaction, document, and insight across petabytes of vector data at low cost, so agents won’t be forced to forget valuable context. Support continual learning, historical context, retraining, and fine-tuning to drive deeper agent intelligence. Whether for agent memory or similarity search across massive AI datasets, S3 Vectors provides a cost-effective data foundation for storing and retrieving vectors.
AI-ready storage for development at any scale
Store and quickly access any amount of vector data to jump-start your AI projects. With no infrastructure setup required, S3 Vectors allows you to put your data to work and begin AI development immediately. It is also built to handle demanding storage requirements for sophisticated AI applications. Whether you're building personalization engines, natural language processing systems, or navigating through large code bases, S3 Vectors provides cost-optimized AI-ready storage that scales to meet your needs—accelerating AI innovation at every step, from prototype to production.
Optimize vector search price and performance with Amazon OpenSearch Service
Balance cost and performance by combining the industry-leading economics for scalable vector storage in S3 Vectors and the high-performance search capabilities of Amazon OpenSearch Service for high-throughput, low-latency vector search. Use S3 Vectors with Amazon OpenSearch Service to lower storage costs for infrequent queried vectors, and then quickly move them to OpenSearch as demands increase or to enhance search capabilities. This strategic integration allows you to allocate vector workloads to the most appropriate service based on performance requirements, ensuring both cost optimization and exceptional query responsiveness.
Customers
March Networks
March Networks, a Delta Group company, works with some of the world’s largest banks and retailers, delivering secure, cloud-based intelligent video solutions, enhancing security, operational efficiency, and profitability through real-time business insights.
“Amazon S3 Vectors delivers clear advantages for large-scale video and photo intelligence. Its cost-optimized architecture allows us to store billions of vector embeddings economically, while seamless integrations with Amazon Bedrock and S3 streamline our gen AI and video workflows. By leveraging S3’s massive scale and eleven nines of durability, we gain the stability required to manage ever-growing volumes of video data and vector embeddings. With high-throughput and low-latency semantic search, we can achieve sub-second insights across entire video archives. S3 Vectors provides the scalable, cost-effective storage layer essential for massive photo and video analytics at scale.”
Jeff Corrall, Chief Product Officer, March Networks
Qlik
Qlik is a global software company in AI-powered data analytics and integration, enabling organizations to make faster, more informed decisions through real-time data access and insights. Its end-to-end platform combines AI, automation, and governed data workflows to transform raw data into actionable intelligence.
“We ingested hundreds of million vectors supported by a large number of resource indexes, leveraging S3 vectors engine fronted by OpenSearch. This will enable a complete semantic search functionality across all entities in our analytics and data Integration products for data engineers, analytic consumers and AI agents alike.”
Martin Andersson, Chief Architect, Qlik
MIXI
MIXI, Inc. delivers social communication and digital entertainment experiences at scale, reaching millions of users through mobile gaming, sports engagement, and community platforms. By combining deep customer understanding with data-driven innovation, MIXI builds interactive services that connect people and enrich everyday life.
"By adopting Amazon S3 Vectors, we're able to build flexible, metadata-aware semantic search capabilities that scale to serve our FamilyAlbum photo-sharing community of more than 27 million users. The fully managed infrastructure greatly simplifies operations compared to self-managed search systems, allowing our team to focus on delivering new AI-powered features. With plans to index roughly 400 million vectors across 100 indexes, S3 Vectors gives us the performance and cost efficiency we need to expand semantic search, powering future experiences like personalized photo print recommendations for every user."
Takahiro Kinouchi, ML Engineer, MIXI, Inc.
Backlight
Backlight is a global media technology company that replaces broken media workflows with simple, AI-powered products. Through its integrated suite of solutions, Backlight empowers creative and production teams to focus on crafting compelling stories that drive impact.
"We have hundreds of customers with 1,000 hour plus video libraries, some in the hundreds of thousands. They need to make intelligent decisions on distributing their content to their owned and operated free ad-supported streaming television (FAST) and apps audiences. Amazon S3 Vectors gives us the foundation to scale intelligent media workflows, allowing our customers to enrich their media with searchable data across the largest libraries."
Ed Laczynski, GM, Zype, Backlight, Backlight
Twilio
Twilio enables companies to use communications and data to add intelligence and security to every step of the customer journey. Today’s leading companies trust Twilio to build direct, personalized relationships with their customers.
“S3 Vectors puts an accessible vector interface right inside the storage we already trust, giving us the scale of S3 with the intelligence of semantic search in a single click. That simplicity lets Twilio teams plug powerful retrieval augmented generation and personalized recommendations into our customer engagement platform without new infrastructure or tuning headaches. We’re excited to see how S3 Vectors helps developers turn everyday data into smarter, more trusted customer experiences.”
Zachary Hanif, Head of AI, ML, and Data; VP of Traffic Intelligence, Twilio
TwelveLabs
TwelveLabs is a pioneer in multimodal AI, specializing in advanced video understanding technology. Its video foundation models enable organizations to search, summarize, and analyze their video content with human-like precision—by understanding not just what’s visible on screen, but the rich context and meaning behind it.
“Video holds some of the most valuable and underutilized information in the world, but until now it’s been locked behind time-consuming manual workflows. Our foundation models enable our customers to turn petabytes of video into searchable, actionable knowledge. With scalable infrastructure like Amazon S3 Vectors, we can deliver semantic search and video analyzation at enterprise scale—empowering teams to focus on creativity, decision-making, and impact.”
Jae Lee, Co-Founder & CEO, TwelveLabs
Spice AI
Spice AI helps enterprises build fast, accurate, and scalable AI applications and
agents using its portable, open-source data and AI compute engine. It unifies
data and search from disparate sources and supports workloads across cloud,
edge, and on-premises systems, simplifying AI development.
“The industry is increasingly relying on object storage as AI applications and agents need access to growing data volumes. Amazon S3 Vectors is incredibly exciting, as we can now get S3 scale, price point, elasticity, and durability in a simple solution for semantic search and retrieval. We’ve partnered with the S3 team to integrate S3 Vectors into the Spice.ai open-source data and AI compute engine, offering a simple SQL interface to efficiently manage and query vector embeddings across enterprise data sources.”
Luke Kim, Founder and CEO, Spice AI
xCures
xCures operates an AI-assisted healthcare data platform that extracts clinical information from aggregated, structured, and normalized medical records.
“S3 Vectors provides a cost-effective complement to Amazon OpenSearch Service for vector management, helping us scale efficiently while maintaining the performance requirements we need for distinct workloads. This enables us to better identify meaningful clinical content in medical records and support high-quality structured data extraction at scale.”
Zach Kaufman, VP of Product Management, xCures
BMW
The BMW Group is the world’s leading provider of premium cars and motorcycles and the home of the BMW, MINI, Rolls-Royce and BMW Motorrad brands.
“Cloud Data Hub is the central data platform of the BMW Group, managing BMW's curated and extensive datasets stored on S3 with Apache Iceberg. To boost AI-based data usability across the organization, a hybrid search solution is being developed to integrate BMW's structured iceberg data with now also semi-structured column data. S3 Vectors was selected for its optimal balance of cost and performance, as well as its compatibility with the existing S3 Iceberg architecture and identity and access management framework.”
Ruben Simon, Head of Product Management, Cloud Data Hub, BMW
Precisely
Precisely is the trusted partner for data integrity, with decades of deep domain expertise that span software, data and data strategy services. Its portfolio helps integrate customer data, improve data quality, govern data usage, geocode and analyze location data, and enrich it with complementary datasets for confident business decisions.
“We are excited to explore the potential of Amazon S3 Vectors to bring cost-performance flexibility to our AI-powered data discovery and metadata curation capabilities.”
Tendu Yogurtcu, Chief Technology Officer, Precisely
Nomad Media
Nomad Media provides a cloud-native content and asset management, content distribution, and live streaming platform built on AWS that seamlessly merges cloud-based asset management with the power of AWS Media Services and AI/GenAI into one unified easy-to-use system.
“Amazon S3 Vectors has allowed us to efficiently and cost effectively scale our media search capabilities to billions of records for our customer’s ever growing content libraries.”
Adam Miller, Co-founder and CEO, Nomad Media
Natera
Natera specializes in genetic testing using non-invasive, cell-free DNA technology with a focus on oncology, women’s health, and organ health. Doctors and clinics use Natera’s tests to design treatment plans and provide precision medicine to patients.
“We use S3 Vectors and Amazon Bedrock to create vector indexes and ingest vectors for the lab equipment engineering documentation use case. This integration enables our lab equipment service engineers to quickly locate and connect information across complex instrument manuals, significantly improving the speed and accuracy for maintenance and troubleshooting. As a result, Natera achieves faster issue resolution and greater instrument uptime across our laboratory operations.”
Ariel Jirau, Sr. Principal Software Engineer, Natera
Squiz
Squiz, a global Digital Experience Platform provider, uses Amazon S3 Vectors to power its Conversational Search tool, enabling organizations to deliver more engaging website experiences through the natural-language interactions users now expect.
“S3 Vectors has allowed us to reimagine our ingestion pipeline. It increased our conversational data processing speed by 50% and reduced costs by letting us shift from bespoke, always-on infrastructure to a scalable serverless model. We can now scale seamlessly from 25,000 to millions of vectors per client, allowing our engineering teams to focus on RAG innovation instead of infrastructure management.”
Greg Sherwood, CTO, Squiz
Did you find what you were looking for today?
Let us know so we can improve the quality of the content on our pages