Amazon Web Services did not begin as a grand plan to remake the technology industry. It began as an answer to a frustration.
In the early 2000s, Amazon was straining under the weight of its own growth. Every new project, whether a recommendation engine, a checkout flow, or a partner integration, seemed to require engineers to reinvent the same plumbing: storage, compute, messaging, databases. Teams duplicated effort, stepped on each other’s code, and shipped slowly.
Around 2002, Jeff Bezos issued what became famous inside Amazon as the “API mandate.” Every team had to expose its data and functionality through service interfaces, communicate only through those interfaces, and design them as if they would one day be exposed to outside developers. The penalty for not complying was termination. The directive forced Amazon to rebuild itself as a collection of loosely coupled services.
That internal discipline became the seed of an external business. In 2003, engineers Benjamin Black and Chris Pinkham wrote a short paper imagining a standardized, automated infrastructure Amazon could sell as a utility, pay as you go, like electricity. Bezos liked it. Andy Jassy, then his technical chief of staff, was tasked with building the business.
The first services rolled out quietly. Simple Queue Service launched in 2004. In March 2006, Amazon released S3, which let developers store any amount of data over the web. Five months later came EC2, which let anyone with a credit card rent virtual servers by the hour. For the first time, a startup founder could spin up real computing capacity in minutes instead of buying hardware and waiting weeks.
The industry was slow to understand. Analysts wondered why a bookseller was renting servers. But developers understood, and they came in waves. Within a decade, AWS was running large portions of the internet and generating most of Amazon’s operating profit.
What began as a memo about internal APIs ended up reshaping how software is built, deployed, and paid for.