Amazon Web Services: Cloud, AI, and Bedrock | hokai.io

Amazon Web Services (AWS), founded 2006, leads global cloud with 30% market share and $128.7B FY2025 revenue. Bedrock, SageMaker, and Trainium power AI.

Amazon Web Services (AWS), launched 2006 and headquartered in Seattle, is the world's largest cloud platform with approximately 30% global market share and $128.73 billion in FY2025 revenue. CEO Matt Garman leads 200+ cloud services including Amazon Bedrock (AI model access), SageMaker (ML development), and Trainium (custom AI silicon). Amazon has committed $200 billion in 2026 capex, primarily for AWS expansion.

Amazon Web Services (AWS) launched in 2006 from Seattle, Washington as a division of Amazon.com. It is the world's largest cloud infrastructure provider, holding approximately 30% global market share, with FY2025 revenue of $128.73 billion and Q1 2026 revenue of $37.59 billion (28% growth). AWS CEO Matt Garman leads the division. Key AI products include Amazon Bedrock, SageMaker, and the Trainium custom AI chip family.

Founded: 2006 · HQ: Seattle, WA, USA · Team: 100000-150000 · CEO: Matt Garman · Funding: Division of Amazon.com (NASDAQ: AMZN). FY2025 revenue $128.73B. Amazon capex $200B committed for 2026, primarily AWS expansion.

About Amazon Web Services

Amazon Web Services launched publicly in 2006 with the release of Amazon S3 (object storage) and Amazon EC2 (compute), services conceived by Andy Jassy and a small team at Amazon who recognized that internet companies were rebuilding the same undifferentiated infrastructure from scratch. The thesis was simple: rent computing resources on demand rather than buying hardware upfront. AWS incorporated as a distinct Amazon business unit and has been headquartered in Seattle, Washington, alongside its parent company Amazon.com. Over two decades, AWS grew from a side project to the engine that generates most of Amazon's operating income, reporting $128.73 billion in revenue for full-year 2025. AWS operates more than 200 fully featured services spanning compute, storage, databases, networking, machine learning, analytics, security, and developer tools. Core infrastructure services include Amazon EC2 (virtual machines), Amazon S3 (object storage), Amazon RDS and Aurora (relational databases), Amazon Lambda (serverless compute), and Amazon VPC (private networking). AI and machine learning services include Amazon Bedrock (fully managed access to foundation models from Anthropic, Meta, Mistral, Google, OpenAI, and others), Amazon SageMaker AI (model development and deployment), and Amazon Q (enterprise AI assistant). Custom silicon includes Trainium chips for model training and Graviton chips for general compute, offering 20 to 40 percent better price-performance than third-party CPUs. At AWS re:Invent in December 2025, the company launched Trainium3 UltraServers, packing 144 Trainium3 chips into a single system and delivering 4.4x more compute performance than Trainium2. Trainium4 was previewed with broad availability expected by late 2026. Amazon Bedrock expanded its model catalog to include Google, Kimi AI, MiniMax AI, Mistral AI, NVIDIA, OpenAI, and Qwen models. Amazon Bedrock added reinforcement fine-tuning, showing 66% accuracy gains over base models. In April 2026, AWS announced Bedrock Managed Agents powered by OpenAI (limited preview) and confirmed Anthropic is now training its most advanced models on AWS Trainium and Graviton chips. AWS AI Factories launched to bring dedicated AWS AI infrastructure into enterprise and government data centers. In May 2026, AWS announced AI agent stablecoin payment infrastructure in partnership with Coinbase and Stripe. AWS is a wholly owned division of Amazon.com, Inc. (NASDAQ: AMZN), which does not break out AWS funding separately. Amazon's parent company carries a market capitalization approaching $2 trillion and has committed $200 billion in capital expenditure for 2026, primarily directed at AWS data center expansion and AI infrastructure. AWS revenue grew from $107.56 billion in 2024 to $128.73 billion in FY2025, with Q1 2026 revenue of $37.59 billion representing 28% year-over-year growth. Operating income for Q1 2026 was $12.47 billion. Amazon has invested $25 billion in Anthropic and expanded OpenAI's AWS compute commitment to $100 billion over eight years. AWS earns revenue through a pay-as-you-go consumption model: customers pay by the hour or second for compute, per gigabyte for storage, per request for managed services, and per million tokens for Bedrock AI inference. Enterprise customers negotiate Reserved Instances or Savings Plans for predictable discounts of 40 to 60 percent versus on-demand pricing. AWS Marketplace provides 42,240 third-party software products. The AI segment of AWS alone runs at an annualized revenue rate exceeding $15 billion as of early 2026, nearly 260 times larger than the entire company's revenue run rate three years after its original 2006 launch. AWS CEO Matt Garman joined AWS as an intern in 2006, managed Amazon EC2 and Compute Services for over a decade, and was appointed CEO of AWS in May 2024. He reports to Amazon CEO Andy Jassy, who led AWS from 2003 to 2021 before succeeding Jeff Bezos as Amazon's CEO. AWS employs approximately 127,000 workers directly in engineering, sales, and support roles. In early 2026, Amazon eliminated approximately 30,000 roles across its corporate workforce while planning to hire 11,000 software engineers and developers, with AWS managerial layers specifically targeted in the restructuring. AWS data centers span over 30 geographic regions globally, with more than 100 Availability Zones across North America, Europe, Asia-Pacific, South America, the Middle East, and Africa. AWS's stated mission is to enable any organization, from two-person startups to the world's largest governments, to access on-demand compute, storage, and AI services with no upfront capital commitment. The company's long-term research priorities center on custom silicon (Trainium, Graviton, Inferentia), agentic AI infrastructure through AgentCore and Strands (downloaded more than 25 million times), and sovereign AI deployments for regulated industries via AI Factories. AWS does not publish a standalone AI safety policy but is a strategic investor in Anthropic and distributes Anthropic's Claude models through Amazon Bedrock under commercial agreements. AWS holds approximately 29 to 30 percent global cloud infrastructure market share, ahead of Microsoft Azure at 20 percent and Google Cloud at 13 percent, according to Synergy Research Group data from Q3 2025. AWS wins on breadth (200+ services vs. Azure's approximately 170 and GCP's approximately 150), depth of AI chip investment (Trainium3 production, Trainium4 in preview), and enterprise trust built over 20 years. AWS loses on Azure's deep Microsoft integration for enterprise buyers who run Windows Server, Active Directory, and Office 365, and on Google's TPU advantage for specific ML research workloads. Among the 80 percent of Fortune 100 companies using Amazon Bedrock, AWS's distribution advantage is substantial. AWS operates under FedRAMP High authorization for US government workloads, making it the preferred cloud provider for US federal agencies including the CIA, DoD, and numerous civilian agencies. EU GDPR compliance is maintained through Standard Contractual Clauses, Data Processing Addenda, and the availability of EU-resident data center options in Ireland and Frankfurt. AWS complies with ITAR, FISMA, CMMC, DoD IL2 through IL6, and industry-specific frameworks including HIPAA, PCI DSS, and HITRUST CSF. The EU AI Act's general-purpose AI obligations apply to model providers distributing through Bedrock, with AWS operating as an intermediary infrastructure provider rather than a model developer for most Bedrock-hosted models.

Mission

Enable any organization to access reliable, scalable, and inexpensive cloud computing infrastructure and AI services on demand.

Products

  • Amazon EC2 + S3 + Core Infrastructure (Compute, storage, and networking cloud infrastructure): https://aws.amazon.com/products/
  • Amazon Bedrock (Fully managed foundation model access (Anthropic, OpenAI, Google, Mistral, Meta)): https://aws.amazon.com/bedrock/
  • Amazon SageMaker AI (Managed ML model development, training, and deployment platform): https://aws.amazon.com/sagemaker/
  • AWS Trainium + Graviton (Custom AI training and general compute silicon): https://aws.amazon.com/machine-learning/trainium/
  • Amazon Q (Enterprise AI assistant for code, data, and business workflows): https://aws.amazon.com/q/
  • AWS AI Factories (On-premises sovereign AI infrastructure for enterprises and governments): https://aws.amazon.com/solutions/ai-factories/

Compliance

SOC 2 Type II, SOC 3, ISO 27001:2022, ISO 27017:2015, ISO 27018:2019, ISO 27701:2019, FedRAMP High, PCI DSS, HIPAA-eligible, HITRUST CSF, CMMC, DoD IL2-IL6, FIPS 140-3, CSA STAR Level 2

Links

Website · GitHub · Twitter · LinkedIn · Blog · Docs

Frequently Asked Questions

What is Amazon Web Services and what do they build?

Amazon Web Services (AWS) is the cloud computing division of Amazon.com, launched publicly in 2006 from Seattle, Washington with Amazon S3 (object storage) and Amazon EC2 (virtual machines). It is the world's largest cloud infrastructure provider, holding approximately 29 to 30 percent global market share, with FY2025 revenue of $128.73 billion and over 200 fully featured services. AWS offers compute (EC2, Lambda, Fargate), storage (S3, EBS, Glacier), managed databases (RDS, Aurora, DynamoDB), networking (VPC, CloudFront, Direct Connect), AI and ML services (Bedrock, SageMaker, Q), and custom silicon (Trainium for training, Graviton for compute, Inferentia for inference). Amazon Bedrock provides managed access to foundation models from Anthropic, Meta, Mistral, Google, OpenAI, and others. AWS serves over 4 million customers globally, with approximately 80 percent of Fortune 100 companies using Amazon Bedrock. Products are available through pay-as-you-go consumption, Reserved Instance contracts, and the AWS Marketplace, which offers 42,240 third-party software products.

Who founded Amazon Web Services and who is the CEO?

AWS was conceived and founded within Amazon.com by Andy Jassy starting in 2003, launching publicly in 2006 with S3 and EC2. Jassy served as SVP and CEO of AWS from 2003 until July 2021, when he succeeded Jeff Bezos as CEO of Amazon.com. Jeff Bezos, who founded Amazon in 1994, remains Executive Chair of Amazon. Matt Garman, who joined AWS as an intern in 2006 and spent over a decade managing Amazon EC2 and Compute Services, was appointed AWS CEO in May 2024. Garman sits on Amazon's senior leadership team (S-Team) and has overseen the division's rapid AI expansion. Recent S-Team additions include Prasad Kalyanaraman (VP of AWS Infrastructure Services, April 2026) and Dave Brown (promoted to SVP leading EC2, Bedrock, and SageMaker). AWS does not have independent founders outside the Amazon corporate structure; it was built entirely as an Amazon internal project.

How much funding has Amazon Web Services raised?

AWS is a wholly owned division of Amazon.com, Inc. (NASDAQ: AMZN) and does not raise independent funding. Amazon committed $200 billion in capital expenditure for 2026, primarily directed at AWS data center expansion, GPU procurement, and custom chip development. AWS full-year 2025 revenue was $128.73 billion, up from $107.56 billion in 2024, with Q1 2026 revenue of $37.59 billion representing 28% year-over-year growth. AWS operating income was $12.47 billion in Q1 2026, representing the majority of Amazon's total company operating income. Amazon has invested $25 billion in Anthropic and expanded OpenAI's AWS compute commitment to $100 billion over eight years, reflecting the scale of AWS's AI ecosystem bets. Amazon's total market capitalization approaches $2 trillion as of May 2026. AWS carries more than $200 billion in customer revenue backlog and over $225 billion in Trainium chip reservation commitments from enterprise customers.

What products does Amazon Web Services make?

AWS offers more than 200 fully featured cloud services. Core infrastructure includes Amazon EC2 (virtual machines in dozens of instance types), Amazon S3 (object storage at exabyte scale), Amazon Lambda (serverless compute), Amazon RDS and Aurora (managed relational databases), Amazon DynamoDB (NoSQL database), Amazon Redshift (data warehouse), and Amazon VPC (private networking). AI and ML services include Amazon Bedrock (access to frontier models from Anthropic, Meta, Mistral, Google, OpenAI, and Qwen), Amazon SageMaker AI (model training, fine-tuning, and deployment), Amazon Q (enterprise AI assistant for code and data tasks), and Amazon Rekognition, Polly, and Transcribe for vision, speech, and language tasks. Custom silicon includes Trainium3 (training at 4.4x the performance of Trainium2), Graviton4 (general compute at 20-40% better price-performance than x86), and Inferentia (inference optimization). AWS AI Factories brings dedicated sovereign AI infrastructure into enterprise and government premises. Amazon Nova Forge enables custom foundation model training for approximately $100K annually. Pricing is consumption-based with Reserved Instance and Savings Plan discounts of 40 to 60 percent.

Where is Amazon Web Services headquartered and how big is the team?

AWS is headquartered in Seattle, Washington, co-located with Amazon.com's corporate campus. AWS employs approximately 127,000 workers directly in engineering, sales, and support roles, within a broader Amazon corporate workforce of 1.576 million employees as of end-2025. In early 2026, Amazon eliminated approximately 30,000 corporate roles (including AWS management layers) while announcing plans to hire 11,000 software engineers and developers in 2026, reflecting a shift toward leaner, more AI-focused headcount. AWS data centers span more than 30 geographic regions globally, including major clusters in Northern Virginia, Oregon, Ireland, Frankfurt, Tokyo, Singapore, Sydney, and Sao Paulo, with over 100 Availability Zones total. AWS also announced new AI Factories deployments in sovereign data centers for regulated enterprise and government clients. The AWS workforce is distributed globally with major engineering hubs in Seattle, Dublin, Sydney, and Bangalore.

What is Amazon Web Services' mission or research focus?

AWS's mission is to enable any organization, from two-person startups to the world's largest governments, to access on-demand compute, storage, and AI services without upfront capital commitment. The company's research agenda centers on three areas: custom silicon (Trainium for AI training, Graviton for general compute, Inferentia for inference), agentic AI infrastructure (AgentCore for deploying and managing AI agents, Strands framework downloaded 25 million times), and sovereign AI (AI Factories for regulated on-premises deployments). AWS collaborates closely with Anthropic as its primary AI research partner, with Anthropic training its most advanced models on AWS Trainium and Graviton chips. AWS does not publish a standalone AI safety policy or Responsible Scaling Policy but distributes Claude models through Bedrock under commercial agreements with Anthropic. Open source contributions include Graviton compiler toolchains, ML framework optimizations for PyTorch and JAX, and the Strands agent development framework. AWS publishes model cards and documentation for Amazon Nova and Titan models.

Is Amazon Web Services compliant with SOC 2, GDPR, FedRAMP, and HIPAA?

AWS maintains one of the most comprehensive compliance portfolios in the cloud industry, independently audited by third parties and accessible through AWS Artifact. Certifications include SOC 2 Type II, SOC 3, ISO 27001:2022, ISO 27017:2015, ISO 27018:2019, ISO 27701:2019, FedRAMP High (for US government workloads), PCI DSS, HIPAA-eligible services (BAA available), HITRUST CSF, CMMC, DoD Impact Levels 2 through 6, FIPS 140-3, and CSA STAR Level 2. AWS is the cloud provider for numerous US federal agencies including the CIA (C2S contract), DoD, and multiple civilian agencies under FedRAMP High authorization. GDPR compliance is maintained through Standard Contractual Clauses, Data Processing Addenda, and EU-resident data options in Ireland and Frankfurt. AWS operates under a Shared Responsibility Model: AWS secures the infrastructure; customers are responsible for securing their workloads and data within it. Compliance documentation including SOC reports, penetration test summaries, and ISO certificates is available on-demand through AWS Artifact for enterprise customers.

Who are Amazon Web Services' main competitors?

AWS's primary cloud infrastructure competitors are Microsoft Azure and Google Cloud. Against Azure, AWS holds approximately 29 to 30 percent global market share versus Azure's 20 percent, but Azure's deep integration with Microsoft enterprise software (Active Directory, Office 365, GitHub, Teams) creates structural advantages in Windows-heavy enterprise accounts that AWS cannot easily replicate. Against Google Cloud, AWS holds a substantial market share lead, but Google's TPU infrastructure and TensorFlow ecosystem give it advantages in specific ML research and training workloads. Against Oracle Cloud, AWS competes primarily in database migration workloads where Oracle's autonomous database and licensing agreements create incumbent advantages. AWS AI chip competition includes NVIDIA, which dominates the GPU training market globally and sees Trainium as a credible threat to its H100 and B200 dominance. In developer platform competition, Cloudflare's edge compute and Vercel's frontend deployment tools are taking specific developer workloads from AWS Lambda and CloudFront at the margins. AWS's primary advantage against all competitors is its 200+ service breadth, 20+ year enterprise relationships, and the scale of its 2026 infrastructure investment commitments.