Cloud Computing Tutorial · Cloud Basics · Lesson 5 of 24
Cloud Computing History & Milestones
How Amazon, Microsoft and Google turned their infrastructure into public clouds, the launches that defined each era from S3 and EC2 to serverless, Kubernetes and AI accelerators, and how the three providers compare today.
How cloud computing started
Renting computing power is an old idea (1960s time-sharing), but the modern cloud began when Amazon exposed its internal infrastructure as self-service APIs. Amazon Web Services launched SQS (in beta) in 2004, then S3 on 14 March 2006 and EC2 in August 2006: storage and servers by the hour, created with an API call, with no contract.
Google launched App Engine in 2008, a platform for running web apps without managing servers. Microsoft announced Windows Azure in 2008 and made it generally available in February 2010; it was renamed Microsoft Azure in 2014. Google's IaaS, Compute Engine, became generally available in December 2013.
Timeline of milestones
- GravitonHistoric
- AWS's own Arm-based server CPUs; custom silicon becomes a cloud differentiator (later Google Axion and Azure Cobalt)
- EKS and AKSHistoric
- Managed Kubernetes reaches general availability on AWS and Azure
- Nitro SystemHistoric
- AWS moves virtualisation onto dedicated hardware cards, the basis of modern EC2 instances
- GKEHistoric
- Google Kubernetes Engine (then Container Engine) becomes generally available, soon after Kubernetes 1.0
- AWS LambdaHistoric
- Announced at re:Invent; serverless functions billed per request and duration
- Microsoft AzureHistoric
- Windows Azure is renamed, reflecting broad support for Linux and open source
- Compute Engine GAHistoric
- Google Cloud's virtual machines become generally available
- Windows Azure GAHistoric
- Microsoft's cloud platform becomes generally available
- Google App EngineHistoric
- A platform-as-a-service for web applications
- Amazon EC2Historic
- Virtual servers on demand, paid by the hour
- Amazon S3Historic
- Object storage by API, the first widely used AWS infrastructure service
- Amazon SQS betaHistoric
- The first AWS infrastructure service: a hosted message queue
Turning points
- Pay-as-you-go infrastructure (2006): start-ups no longer needed to buy servers before they had customers.
- Managed services (2009 onwards): managed databases (RDS, 2009), queues and caches moved operational work to providers.
- Containers and Kubernetes (2014-2018): portable workloads and managed orchestration on every cloud.
- Serverless (2014 onwards): functions and later serverless containers (Fargate, Cloud Run) billed per use.
- Custom silicon (2017 onwards): Nitro, Graviton, TPUs, Trainium and Inferentia make price-performance a hardware race.
- AI platforms (2020s): GPUs, AI accelerators and managed model services (Amazon Bedrock, Azure OpenAI and AI Foundry, Vertex AI) became central to cloud strategy.
The three providers compared
| AWS | Microsoft Azure | Google Cloud | |
|---|---|---|---|
| Launched (infrastructure) | 2006 | 2010 | 2008 (App Engine), 2013 (Compute Engine) |
| Strengths | Breadth and depth of services, largest ecosystem | Microsoft integration (Entra ID, Windows, .NET, Microsoft 365), enterprise and hybrid | Data and analytics (BigQuery), Kubernetes heritage, global network, AI |
| Account structure | Organizations, accounts | Management groups, subscriptions, resource groups | Organization, folders, projects |
| Infrastructure as code | CloudFormation, CDK (and Terraform) | ARM templates, Bicep (and Terraform) | Infrastructure Manager (Terraform) |
Which cloud should you learn?
- Learn concepts (identity, networking, compute, storage, databases, cost, reliability); they transfer.
- Start with the cloud your employer or target employers use; AWS has the largest share of job postings, Azure is strong in enterprises using Microsoft products, Google Cloud in data and AI-heavy companies.
- Use Terraform to keep skills portable across providers.
Official resources
- AWS documentation and the AWS Well-Architected Framework
- Microsoft Azure documentation and the Azure Well-Architected Framework
- Google Cloud documentation and the Google Cloud Architecture Framework
- Our history (AWS)
Last reviewed 21 October 2026. Release and support dates come from the official release pages linked above.