The Complete History of Cloud Computing: From Time-Sharing to AI Infrastructure

Quick facts

• Cloud computing grew from earlier ideas including time-sharing, utility computing, virtualisation and networked services.
• Salesforce popularised browser-delivered enterprise software after launching in 1999.
• Amazon S3 and EC2 launched in 2006 and made storage and compute available on demand.
• Google App Engine appeared in 2008 and Microsoft Azure became commercially available in 2010.
• NIST defined cloud computing through characteristics such as on-demand self-service, broad network access, resource pooling, rapid elasticity and measured service.

Time-sharing and utility-computing ideas

In the 1950s and 1960s, computers were extremely expensive. Time-sharing systems allowed many users to interact with one mainframe. Researchers and business leaders imagined computing delivered like electricity: customers would consume resources without owning every machine.

These ideas lacked today’s global networks and automation, but they established the economic logic of shared infrastructure.

Virtualisation makes hardware flexible

IBM developed influential virtual-machine systems for mainframes in the 1960s and 1970s. Virtualisation allowed one physical computer to behave like several isolated machines. This improved utilisation and made workloads easier to manage.

Later x86 virtualisation from VMware and others brought similar techniques to commodity servers. Data centres could consolidate applications and provision systems more quickly.

The Internet and software as a service

The commercial Internet allowed applications to be delivered through browsers. Salesforce, founded in 1999, promoted the idea that businesses could use customer-management software without installing and maintaining it locally.

Application service providers existed earlier, but broadband, web standards and subscription pricing made software as a service practical at larger scale.

AWS creates on-demand infrastructure

Amazon’s retail business required teams to build reliable internal infrastructure. The company concluded that developers everywhere faced similar problems. Amazon Web Services launched Simple Storage Service, or S3, in March 2006. Elastic Compute Cloud, or EC2, followed later that year.

Developers could rent storage and virtual servers through APIs and pay for usage. This reduced the time and capital required to launch an Internet business and made infrastructure programmable.

Google and Microsoft enter the platform market

Google launched App Engine in 2008, offering a managed environment for web applications. Microsoft announced Azure in 2008 and made it commercially available in 2010. The major cloud providers developed compute, storage, databases, networking, analytics and identity services.

Competition shifted from renting virtual machines to providing complete platforms for application development and data processing.

Cloud service and deployment models

Infrastructure as a Service supplies virtualised compute, storage and networks. Platform as a Service manages more of the application runtime. Software as a Service delivers complete applications. Serverless services execute code or workflows without customers managing individual servers.

Public cloud uses shared provider infrastructure. Private cloud applies cloud-style automation to dedicated environments. Hybrid and multi-cloud strategies connect multiple locations or providers.

Containers, Kubernetes and cloud-native design

Containers package applications with dependencies while sharing the host operating-system kernel. Docker helped popularise containers in the 2010s. Google released Kubernetes as open source in 2014 to automate deployment and scaling of containerised applications.

Cloud-native systems use APIs, automation, microservices, observability and infrastructure as code. These methods can improve deployment speed but also increase operational complexity.

Serverless, edge and AI clouds

AWS Lambda, introduced in 2014, helped popularise event-driven serverless computing. Providers later added managed databases, data warehouses, machine-learning platforms and specialised chips.

AI training and inference require large clusters of GPUs, high-speed networking and storage. Cloud providers now operate “AI factories” that rent accelerated computing. Edge computing places some processing nearer to users and devices to reduce latency.

Risks and responsibilities

Cloud computing does not remove responsibility. Customers must configure identity, backups, encryption and access correctly. Providers secure the underlying infrastructure, while customers remain responsible for many workload and data choices under a shared-responsibility model.

Other concerns include outages, vendor lock-in, unpredictable costs, data sovereignty, environmental impact and concentration of critical services among a few companies.

Common misconceptions

• “The cloud” is not one place; it is a distributed collection of data centres, networks and software services.
• Cloud does not automatically make an application secure or inexpensive.
• Virtualisation and cloud computing are related but not identical.
• Serverless systems still use servers; customers simply do not manage them directly.

Timeline: key years and locations

Frequently asked questions

Why does this history still matter? Understanding the sequence of inventions, standards, business decisions and public adoption makes it easier to see why today's technology works the way it does. It also separates genuine milestones from popular myths.

Are all dates exact? The article uses specific dates when authoritative sources provide them. Where a technology emerged gradually through research, standardisation and commercial rollout, the text explains the period rather than pretending that a single day created the entire field.

Will this article be updated? Yes. BestAI Newsroom keeps the original publication date and changes the updated date when a correction, newly released archive or important later milestone is added.