AWS
in English: Amazon Web Services
AWS is Amazon's cloud service: you can rent computers, storage, and AI tools from it via the internet, and pay only for what you use.
In plain terms
AWS is like an electricity provider, but instead of power, it offers computing capacity. Just as you do not need to build your own power plant, you do not need to set up and maintain your own server room: you connect to the network and pay for your consumption.
Companies, banks, government agencies, and startups run their websites, applications, and databases on these. Many services you use daily run partially or entirely on AWS machines, though you see none of this.
In the AI era, this is important because training and running large language models requires many expensive, specialized chips. Most companies do not buy these but rent them. On AWS, ready-made AI models and tools for building agents are also available.
It differs from its competitors, Microsoft Azure and Google Cloud, mainly in size and the breadth of its offerings. Of the three major cloud providers, this one started earliest and is also the largest in the market.
How it works
- Amazon operates massive data centers in many parts of the world, containing thousands of servers, storage units, and AI chips.
- Software divides these machines into smaller, isolated units, so the same hardware can be used safely by many clients at once.
- A customer places an order with a few clicks on a website or from a program: for example, ten servers, a database, or access to a language model. These start within minutes.
- The system measures consumption (machine time, stored data, data traffic, and for AI, the volume of processed text) and bills accordingly.
- If traffic increases, the customer activates more capacity; if it decreases, they deactivate it, so they do not pay for unused machines.
An everyday example
You watch a series on a streaming service, order something from an online store, and ask an AI chatbot a question. Quite possibly, AWS machines are doing the work behind all three. If a fault occurs in one of AWS's large regions, meaning a cluster of data centers, several seemingly unrelated apps on your phone can go down at once. This happened, for example, in October 2025, when a disruption in a US AWS region crippled many popular services worldwide.
Who is affected
A common misconception
Limits and risks
- Usage-based billing can easily spiral out of control: a machine accidentally left running or a heavily working AI agent can generate an unexpectedly large bill.
- The more AWS-specific services a company uses, the harder and more expensive it becomes to migrate to another provider later.
- If a major AWS region goes down, services built on it fail en masse, so critical systems require backup solutions.
- Complying with European data protection rules requires extra attention, because the provider is an American company, even if it stores the data in a European data center.
Similar, but not the same
- Amazon (online store) – The Amazon online store sells products to customers, while AWS is a separate business unit of the same company that rents out IT capacity.
- Cloud computing – The cloud itself is the method, meaning renting computing over the internet, while AWS is a specific company offering this.
- Data center – The data center is the physical building with the machines, while AWS is the collection of services built on top of it, available worldwide.
Related concepts
Cloud computing, Cloud service provider, AI infrastructure
Amazon Bedrock, Amazon EC2, Amazon S3, Amazon SageMaker
Microsoft Azure, Google Cloud, Data center, GPU, vendor lock-in
A brief history
Amazon began renting out its own IT expertise developed for its online store to others. AWS launched in its current form in 2006, first with storage services (S3), then with rentable servers (EC2). Amazon's AI service, Bedrock, was made generally available in 2023.
Where things stand
At the end of 2026, AWS is the world's largest cloud provider. Behind it comes Microsoft Azure, followed by Google Cloud with a much smaller share. A large part of AWS investment goes into AI data centers, its own chips (Trainium, Inferentia), and tools for running agents.
In the news
- AWS launches Kiro bots to coordinate AI coding agentsNov 5, 2026 · Code
- AWS launches new package to help bring physical AI into manufacturingNov 5, 2026 · Ipar
- AWS Transform can be used in GitHub Copilot via MCP for .NET projectsOct 27, 2026 · Code
- Omnicom's Credera signs five-year AWS deal for AI marketingOct 20, 2026 · Marketing
- Amazon plans to invest more than €16 billion in Germany in 2027Oct 20, 2026 · Ipar
- AWS energy lead says utilities can use AI to clear grid backlogs; Florida utilities get AI platformOct 15, 2026 · Ipar
- Amazon launches AWS Physical AI Toolchain for roboticsOct 12, 2026 · Ipar
- AWS GovCloud adds support for Claude Code and Claude 5.5 for regulated workloadsOct 5, 2026 · Code
Also known as
AWS · Amazon Web Services · Amazon AWS · Amazon cloud · Amazon's cloud service · Amazon Cloud · AWS cloud · AWS Cloud
Sources
This entry was machine-edited and has undergone machine expert review; the “In the news” section updates automatically from Agent Age stories.