Fees are based on a combination of usage (known as a "Pay-as-you-go" model), hardware, operating system, software, and networking features chosen by the subscriber requiring various degrees of availability, redundancy, security, and service options.

[8] Amazon markets AWS to subscribers as a way of obtaining large-scale computing capacity more quickly and cheaply than building an actual physical server farm.

Clients can interact with these APIs in various ways, including from the AWS console (a website), by using SDKs written in various languages (such as Python, Java, and JavaScript), or by making direct REST calls.

[23] Amazon created "a shared IT platform" so its engineering organizations, which were spending 70% of their time on "undifferentiated heavy-lifting" such as IT and infrastructure problems, could focus on customer-facing innovation instead.

[24][25] Besides, in dealing with unusual peak traffic patterns, especially during the holiday season, by migrating services to commodity Linux hardware and relying on open source software, Amazon's Infrastructure team, led by Tom Killalea,[26] Amazon's first CISO,[27] had already run its data centers and associated services in a "fast, reliable, cheap" way.

[25] By the summer of 2003, Andy Jassy had taken over Bryar's portfolio[31] at Rick Dalzell's behest, after Vermeulen, who was Bezos' first pick, declined the offer.

[22] Jassy subsequently mapped out the vision for an "Internet OS"[15][17][19][32] made up of foundational infrastructure primitives that alleviated key impediments to shipping software applications faster.

[19] Werner Vogels cites Amazon's desire to make the process of "invent, launch, reinvent, relaunch, start over, rinse, repeat" as fast as it could was leading them to break down organizational structures with "two-pizza teams"[c] and application structures with distributed systems;[d] and that these changes ultimately paved way for the formation of AWS[21] and its mission "to expose all of the atomic-level pieces of the Amazon.com platform".

In late 2003, the concept for compute,[e] which would later launch as EC2, was reformulated when Chris Pinkham and Benjamin Black presented a paper internally describing a vision for Amazon's retail computing infrastructure that was completely standardized, completely automated, and would rely extensively on web services for services such as storage and would draw on internal work already underway.

Near the end of their paper, they mentioned the possibility of selling access to virtual servers as a service, proposing the company could generate revenue from the new infrastructure investment.

[59] The three-day event was held in Las Vegas because of its relatively cheaper connectivity with locations across the United States and the rest of the world.

[71] In 2016 Q1, revenue was $2.57 billion with net income of $604 million, a 64% increase over 2015 Q1 that resulted in AWS being more profitable than Amazon's North American retail business for the first time.

[73] Around the same time, Amazon experienced a 42% rise in stock value as a result of increased earnings, of which AWS contributed 56% to corporate profits.

The employee was convicted of hacking into the company's cloud servers to steal customer data and use computer power to mine cryptocurrency.

These models, offered through Amazon Bedrock, are designed for various tasks including content generation, video understanding, and building agentic applications.

[96] In October 2021, it was reported that spy agencies and government departments in the UK such as GCHQ, MI5, MI6, and the Ministry of Defence, have contracted AWS to host their classified materials.

[97] In 2022 Amazon shared a $9 billion contract from the United States Department of Defense for cloud computing with Google, Microsoft, and Oracle.

[7] Each region has multiple "Availability Zones",[113] which consist of one or more discrete data centers, each with redundant power, networking, and connectivity, housed in separate facilities.

[115] As of March 2024,[update] AWS has announced the planned launch of six additional regions in Malaysia, Mexico, New Zealand, Thailand, Saudi Arabia, and the European Union.

[117] In the United States, AWS's partnerships with renewable energy providers include Community Energy of Virginia, to support the US East region;[118] Pattern Development, in January 2015, to construct and operate Amazon Wind Farm Fowler Ridge;[119] Iberdrola Renewables, LLC, in July 2015, to construct and operate Amazon Wind Farm US East; EDP Renewables North America, in November 2015, to construct and operate Amazon Wind Farm US Central;[120] and Tesla Motors, to apply battery storage technology to address power needs in the US West (Northern California) region.

[127] The pop-up lofts in New York[128] and San Francisco[129] are indefinitely closed due to the COVID-19 pandemic while Tokyo has remained open in a limited capacity.

In partnership with the Prince's Trust and the Ministry of Defence (MoD), AWS will help to provide re-training opportunities for young people from disadvantaged backgrounds and former military personnel.

AWS is working alongside a number of partner companies including Cloudreach, Sage Group, EDF Energy, and Tesco Bank.

[132] In 2016, Greenpeace assessed major tech companies—including cloud services providers like AWS, Microsoft, Oracle, Google, IBM, Salesforce and Rackspace—based on their level of "clean energy" usage.

Greenpeace credited AWS for its advances toward greener computing in recent years and its plans to launch multiple wind and solar farms across the United States.

It scanned more than 16.5 million records of naturalized Americans and flagged approximately 124,000 of them for manual analysis and review by USCIS officers regarding denaturalization.

[144] In response to the Log4Shell vulnerability, AWS released hot patch solutions to mitigate risks in Java applications across various environments, including standalone servers, Kubernetes clusters, and Elastic Container Service (ECS).

However, researchers from Unit 42 at Palo Alto Networks identified critical security flaws in these patches that could be exploited for container escape and privilege escalation, potentially granting attackers unauthorized root-level access to the host system.

The issue stemmed from how some users configured ALB's authentication handoff to third-party services, potentially enabling unauthorized access to application data.

This can occur due to various reasons, including but not limited to misconfigurations, security breaches, complex pricing—especially when multiple AWS services are used together—and unexpected data transfer charges.