The Story of Amazon Web Services: Building the Cloud Industry from Scratch

In 2006, a small internal tool built by Amazon engineers to manage their own infrastructure quietly went live for external customers. Nobody predicted that this humble experiment would ignite a trillion-dollar industry—and fundamentally reshape the internet itself.

The Accidental Revolution

Amazon Web Services didn’t start as a grand business plan. It started as a pain point. By the early 2000s, Amazon’s engineering teams were struggling under their own success. Each new feature required months of provisioning servers, managing databases, and fighting with hardware. Developers spent more time babysitting infrastructure than writing code.

A team led by engineer Benjamin Black drafted an internal memo in 2003 that proposed a radical idea: why not build a common infrastructure platform that any Amazon team could use? The idea seemed simple, but its implications were staggering—if Amazon could abstract away the hardware layer just like the operating system abstracts memory, developers could focus purely on building applications.

CEO Jeff Bezos saw the memo and immediately grasped something others missed: if this tool was useful to Amazon engineers, it could be useful to everyone on the internet. He issued the infamous "API mandate"—all teams must communicate internally only through APIs, and those APIs must be designed as if they were public. This forced Amazon to treat its infrastructure as a product, not an internal tool.

The First Building Blocks

AWS launched with two foundational services in 2006:

• Amazon Simple Queue Service (SQS): A message queue that decoupled application components
• Amazon S3 (Simple Storage Service): An infinite, dirt-cheap storage system

These were modest beginnings. S3 offered nothing fancy—just store and retrieve files via REST APIs. But it solved a huge headache. Before S3, storing files required managing physical hard drives, RAID arrays, and backup tapes. S3 made storage an API call away.

The Amazon Elastic Compute Cloud (EC2) followed in 2006 as a private beta. This was the true game-changer. EC2 let users spin up virtual servers in minutes, pay by the hour, and destroy them when done. At that time, provisioning a server typically took weeks—order hardware, wait for shipping, rack it, configure it. EC2 compressed that into a single API call.

The "Get Big Fast" Philosophy

AWS grew through what Bezos called the "flywheel" effect: lower prices attract more customers, more customers drive more utilization, which enables even lower prices. AWS dropped prices over 80 times in its first decade—often unilaterally, without customer requests. Each price cut expanded the potential use cases, pulling in startups, enterprises, and eventually government agencies.

By 2008, AWS was running Netflix’s streaming pipeline. By 2010, it supported Airbnb’s explosive growth. Startups no longer needed to raise millions for servers before launching a product—they could build the entire company on AWS credit.

The Innovation Flywheel

AWS didn’t rest on EC2 and S3. It relentlessly expanded into adjacent spaces:

• 2008: Amazon CloudFront (CDN)
• 2009: Amazon RDS (Relational Database Service), Elastic Load Balancing, Auto Scaling
• 2010: Amazon Route 53 (DNS), Amazon VPC (Virtual Private Cloud)
• 2012: Amazon DynamoDB (NoSQL database), Redshift (data warehouse)
• 2014: AWS Lambda (serverless compute)
• 2017: Amazon SageMaker (machine learning)

Each service abstracted away another slice of undifferentiated heavy lifting. Lambda, in particular, was revolutionary—it let developers run code without managing any servers at all. Just upload code, set a trigger, and AWS handled execution, scaling, and billing by the millisecond.

The Underdog Competitor

By 2015, AWS commanded roughly 30% of the cloud market. Rivals like Microsoft Azure and Google Cloud Platform had poured billions into catching up, but AWS kept its lead through three advantages:

1. Depth: Over 200 services vs competitors’ 60–100
2. Maturity: Years of battle-testing at massive scale
3. Cost leadership: AWS’s margin structure allowed price wars that smaller players couldn’t sustain

The Real Legacy: Changing How Software Gets Built

The most profound impact of AWS isn’t financial. It’s that AWS destroyed the distinction between software and infrastructure. Before AWS, building a product required two parallel businesses: the software business and the infrastructure business. AWS eliminated the second. This single shift unlocked thousands of startups that would never have existed otherwise.

Today, AWS powers critical infrastructure—NASA’s Mars rover data processing, the New York Stock Exchange, the U.S. intelligence community, and entire countries' government services. It runs over a million servers across 30+ geographic regions.

But the deepest legacy is cultural. AWS taught the entire tech industry that hardware doesn’t matter—what matters is what you build on top of it. Every modern cloud provider, every DevOps tool, every container orchestration system descends from the original insight of that 2003 internal memo: infrastructure should be an API call.

In the process, Amazon accidentally built the most profitable business in the history of technology. AWS alone would be a Fortune 500 company. Its revenue in 2023 was over $90 billion—more than Amazon’s entire retail business when AWS first started.

Not bad for an internal tool nobody asked for.