Amazon S3
Amazon Simple Storage Service (S3) launched March 14, 2006, as the first generally available AWS service, and remains the dominant object storage platform globally.
Profile
Stores and retrieves any volume of data objects globally via HTTP, with 11 nines of durability, multiple storage tiers, and native AI integrations.
Amazon Simple Storage Service (S3) launched March 14, 2006, as the first generally available AWS service, and remains the dominant object storage platform globally. Built on an architecture designed for simplicity, durability, and scale, S3 stores 500 trillion objects and 100+ exabytes of data while serving 200 million requests per second. The service pioneered cloud storage with a pay-as-you-go pricing model and has grown to support over 1.167 million companies, from micro-businesses (68% of users have 1–10 employees) to enterprises like Netflix, LinkedIn, and Meta.
AWS maintains roughly 29% market share in cloud infrastructure as of Q1 2026, though this represents a gradual decline from peak dominance in 2022. S3's revenue is not separately disclosed; AWS as a whole is Amazon's fastest-growing business unit, delivering continued double-digit expansion even as competition intensifies from Microsoft Azure and Google Cloud. In 2025–2026, AWS has accelerated S3 product development, introducing S3 Vectors for native vector storage in AI applications, increasing maximum object size from 5TB to 50TB, and rolling out S3 Tables with Apache Iceberg support and intelligent tiering.
Pricing complexity remains a notable friction point—IT leaders report widespread difficulty optimizing S3 bills despite years of adoption, with storage rates varying 23× across storage classes. AWS addressed this via Intelligent-Tiering (saving customers $6B+ cumulatively) and Express One Zone price cuts of 31% for storage and 85% for GET operations in late 2025. The service faces mounting pressure from emerging GPU-focused cloud providers (CoreWeave, Crusoe, Lambda Labs) carving out AI infrastructure niches, and from Azure and GCP capturing disproportionate AI workload growth. Yet S3 remains the reference implementation for object storage and the default choice for data lakes, analytics, and backup across enterprises.
Who buys this
- Analytics teams building data lakes and running SQL queries via Athena and Iceberg
- AI/ML practitioners training models and storing embeddings with S3 Vectors
- Media and entertainment companies archiving and streaming video (BBC, Adobe)
- SaaS and startup backends using S3 for application data, user-generated content, and backups
- Enterprises managing compliance-critical archives with immutable storage and Glacier Deep Archive
Publicly disclosed clients
- Netflix
- Meta
- Adobe
- BBC
- TwelveLabs
- Indeed
Strengths and what to watch
Strengths
- Unmatched scale: 500 trillion objects, 200M requests/second, proven 11-nines durability and 99.99% SLA
- Storage-first architecture for AI: S3 Vectors reduces vector database costs by up to 90% and is natively integrated with Bedrock and OpenSearch
- Mature ecosystem: 1M+ data lakes, 200+ AWS services, global footprint with 32 regions and 102+ AZs, and multi-decade operational track record
Watch for
- Market share erosion: AWS slipped from peak dominance in Q2 2022 and now trails in AI workload adoption; GCP growth (63% YoY in Q1 2026) and Azure momentum (40% YoY) compress S3's competitive moat
- Pricing opacity risks: 94% of IT leaders struggle to optimize S3 costs despite Intelligent-Tiering; Byzantine tiering and transfer pricing create customer churn and switching-cost recovery opportunities for competitors
- Niche fragmentation: CoreWeave, Crusoe, and Lambda Labs are capturing GPU-intensive AI training workloads that historically favored general-purpose cloud; S3 Vectors and custom Trainium/Graviton chips are AWS responses but unproven at scale
Recent moves
- 10mo ago Amazon S3 Vectors reaches general availability with 40x capacity increase to 2 billion vectors per index, 100ms latency, and up to 90% cost reduction vs. specialized vector databases
- 10mo ago AWS re:Invent 2025: S3 batch operations 10x faster, max object size raised to 50TB, S3 Tables gains intelligent-tiering and multi-region replication, FSx for NetApp ONTAP integrates S3 access
- 11mo ago AWS market share slips to 29% as Microsoft Azure (24%) and Google Cloud (12% but growing 63% YoY) gain ground in AI-driven cloud adoption
- 11mo ago S3 Express One Zone storage pricing cut 31%, GET request costs cut 85%, making single-digit millisecond latency viable for cost-sensitive ML and high-frequency workloads
- 1y ago AWS retires legacy free tier for S3 on new accounts, replacing it with $200 credit pool; existing accounts unaffected but signals end of unlimited free-tier era
Key Information
- Industry
- Storage
- Founded
- 2017
- Headquarters
- Seattle, Washington
Frequently Asked Questions
What is Amazon S3?
Amazon S3 is an object storage service that stores and retrieves any volume of data globally via HTTP. It powers 500 trillion objects and serves 200 million requests per second with 11-nines durability. Over 1.167 million companies use S3 for data lakes, backups, archives, and analytics.
How durable is Amazon S3?
S3 guarantees 11-nines of durability, meaning data survives component failures across distributed systems. It also provides a 99.99% SLA and has maintained these standards since 2006 across 32 global regions and 102+ availability zones. S3 stores 100+ exabytes of data reliably.
What is S3 Vectors and when should I use it?
S3 Vectors is native vector storage for AI applications, reaching general availability in December 2025. It reduces vector database costs by up to 90% and supports 2 billion vectors per index with 100ms latency. Integrations with AWS Bedrock and OpenSearch streamline AI workloads and embeddings storage.
How can I reduce my S3 costs?
AWS Intelligent-Tiering saves customers $6 billion cumulatively by automatically moving data to cheaper storage classes. S3 Express One Zone cut storage prices 31% and GET request costs 85% in late 2025. However, 94% of IT leaders still struggle with optimization despite years of adoption due to complex pricing.
What companies use Amazon S3?
Major S3 users include Netflix, LinkedIn, Meta, Adobe, BBC, and Indeed. The service powers over 1.167 million companies, from 68% of small teams with 1–10 employees to large enterprises. These customers use S3 for data lakes, backups, video streaming, and analytics.
Is S3 losing market share to competitors?
AWS market share slipped to 29% as of Q1 2026, down from peak dominance in Q2 2022. Microsoft Azure holds 24% and Google Cloud 12%, but GCP is growing 63% year-over-year in AI workloads. Niche GPU-focused providers like CoreWeave, Crusoe, and Lambda Labs are capturing specialized AI training workloads.
How Amazon S3 compares
Direct head-to-head against 3 competitors. Picked by 7wData.
Amazon S3
- Positioning
- Stores and retrieves any volume of data objects globally via HTTP, with 11 nines of durability, multiple storage tiers, and native AI integrations.
- Customer segments
- Analytics teams building data lakes and running SQL queries via Athena and Iceberg
- Strengths
- Unmatched scale: 500 trillion objects, 200M requests/second, proven 11-nines durability and 99.99% SLA
- Watch for
- Market share erosion: AWS slipped from peak dominance in Q2 2022 and now trails in AI workload adoption; GCP growth (63% YoY in Q1 2026) and Azure momentum (40% YoY) compress S3's competitive moat
- Recent moves
- Amazon S3 Vectors reaches general availability with 40x capacity increase to 2 billion vectors per index, 100ms latency, and up to 90% cost reduction vs. specialized vector databases
Google Cloud Storage
- Positioning
- GCP's AI-workload storage layer, backed by 63% YoY GCP revenue growth in Q1 2026 driven by enterprise AI migration.
- Customer segments
- Analytics and AI/ML teams on GCP, media enterprises, and organizations prioritizing multi-region data durability.
- Strengths
- Storage Intelligence, deployed by 70% of GCP's largest customers, surfaces cross-project cost anomalies and security posture gaps at 50B-object scale.
- Watch for
- Google doubled CDN Interconnect peering egress rates May 1, 2026, compounding cost complaints from multi-CDN enterprise teams.
- Recent moves
- April 2026 (Google Cloud Next 26): Managed Lustre reached GA at 10 TB/s throughput; Cloud Storage MCP server reached GA for AI application integration.
Microsoft Azure Blob Storage
- Positioning
- Unified storage for the enterprise AI lifecycle, embedded in Microsoft stack and powering OpenAI at exabyte scale.
- Customer segments
- Enterprise Microsoft stack teams, AI builders on Azure AI Foundry, and compliance-heavy financial and healthcare organizations.
- Strengths
- Exclusive infrastructure for OpenAI training and serving, natively integrated into Azure AI Foundry and AI Search agent pipelines.
- Watch for
- July 2026 minimum billable object size (128 KiB for Cool/Cold/Archive tiers) raises bills unexpectedly for teams storing millions of small files.
- Recent moves
- April 2026: Smart Tier reached general availability, automating Hot/Cool/Cold transitions with no retrieval fees, after November 2025 Ignite public preview.
Cloudflare R2
- Positioning
- S3-compatible object storage with zero egress fees, targeting workloads where data transfer costs dominate the bill.
- Customer segments
- Developer teams, SaaS backends, AI inference pipelines, and content delivery platforms where data frequently exits storage.
- Strengths
- Zero egress fees on all data leaving R2, eliminating the largest variable cost in object storage for bandwidth-heavy workloads.
- Watch for
- Object Lock (WORM compliance) absent: R2 bucket locks do not satisfy SEC Rule 17a-4 or HIPAA immutable storage mandates, blocking regulated-industry adoption.
- Recent moves
- April 2025: R2 Data Catalog public beta launched, embedding Apache Iceberg tables natively in R2 buckets for zero-egress analytics workloads.
Sources
- aws.amazon.com — S3 product overview, durability guarantees (11 nines), availability SLA, key features, pricing, use cases, customer names (BBC, Adobe, Indeed, GE HealthCare, TwelveLabs)
- www.cloudzero.com — Major AWS/S3 customers (Netflix, Twitch, LinkedIn, Meta, Turner Broadcasting, X, ESPN, NASA, CIA, Samsung), their estimated spend and service usage patterns
- technologychecker.io — 1.167 million companies using S3, customer breakdown by size (68.3% have 1–10 employees), geographic distribution (US 61%, UK 9.81%, Canada 4.77%), top industries (Marketing, Education, Digital Marketing)
- aws.amazon.com — S3 Vectors GA announcement (Dec 2025), 40x scale increase to 2B vectors/index, 100ms latency, 90% cost reduction, 14 regional availability, integrations with Bedrock and OpenSearch
- www.theregister.com — AWS market share 29% (Q3 2025, down from peak in Q2 2022), Azure 20–24%, GCP 12–14% with 63% YoY growth in Q1 2026; big three control 63% of cloud infra spend
- kinsta.com — AWS dominance (one-third of global cloud market), 32 global regions, 102+ AZs, 1.5M+ servers, 200+ services, S3 13.3% share of enterprise data storage market
- aws.amazon.com — re:Invent 2025 storage announcements: S3 Vectors GA, S3 Tables (intelligent-tiering, multi-region replication), batch operations 10x faster, max object 50TB, Storage Lens enhancements, FSx/NetApp ONTAP S3 integration
- blocksandfiles.com — S3 Express One Zone pricing cuts (31% storage, 85% GET), single-digit millisecond latency improvements, 50TB max object size, batch performance gains