Amazon CEO Andy Jassy just shared a rare behind-the-scenes look at AWS's turbulent early days, revealing that even inside Amazon, people thought the cloud venture was "nutty." Speaking to employees this week, Jassy opened up about the doubt, uncertainty, and internal resistance that nearly derailed what would become the company's most profitable division, generating over $90 billion annually.
The confession came during an all-hands meeting where Amazon CEO Andy Jassy dropped his usual executive polish and got brutally honest about AWS's rocky beginnings. "While we were excited about what we were building, I would not say that we were filled with unbridled confidence," Jassy told employees. "There was a lot of uncertainty as we were building. People inside the company thought it was nutty."
The admission offers a fascinating glimpse into what's now become Amazon's cash cow. AWS generated $90.8 billion in revenue last year, representing roughly 70% of the company's operating income. But back in 2006, when Jassy was leading the project, cloud computing was an unproven concept that seemed to distract from Amazon's core e-commerce business.
Jassy's turning point came when his team stopped fixating on outcomes they couldn't predict. "We can't control whether people will use this new form of computing, we can't control how fast it'll grow, we can't control if it's a successful business," he explained. Instead, they zeroed in on what they could influence: "We can control what we define, and we can control what we build, and we can control who we hire."
This mindset shift transformed AWS from an internal experiment into a systematic business-building machine. The team focused on controllable variables like pricing models, customer targeting, and development speed. "Once we started focusing on controlling what we could control, we got to a much better spot," Jassy said.
The strategy worked. AWS launched with simple storage and computing services in 2006, but its focus on developer-friendly tools and aggressive pricing helped it capture early enterprise adoption. By 2015, AWS was generating $7.9 billion annually. Today, it dominates cloud infrastructure with a 32% market share, ahead of Microsoft Azure's 23% and Google Cloud's 11%.
Jassy's reflection on persistence comes at a critical moment for Amazon. The company is pouring billions into AI infrastructure, betting that generative AI will drive the next wave of cloud growth. AWS recently announced new AI chips and expanded partnerships with companies like Anthropic, investments that mirror the uncertain early days of cloud computing.
"I haven't regretted a lot of the decisions I've made in my life," Jassy told employees. "But the ones I've looked back on and felt a little bit of remorse about have been the decisions where I left something where I felt like I really didn't see it through." He emphasized that his biggest regrets weren't failed projects, but promising ventures he abandoned when things got difficult.
The message resonates beyond Amazon. In an era where startup funding has tightened and tech companies face pressure for quick returns, Jassy's emphasis on persistence through uncertainty offers a counternarrative. "It's very rare that you catch lightning in a bottle, where you have an idea, you launch it a few days and it's a home run," he said.
For AWS, the patience paid off spectacularly. What started as an internal skepticism magnet now powers everything from Netflix's streaming infrastructure to the CIA's classified operations. The business unit employs over 100,000 people and continues expanding into edge computing, AI services, and quantum computing research.
Jassy's willingness to discuss AWS's vulnerable early days signals something broader about Amazon's current strategy. As the company navigates new uncertainties around AI regulation, antitrust scrutiny, and slowing e-commerce growth, the CEO seems to be reminding employees that breakthrough innovations require staying in the boat when the waters get choppy.
Jassy's candid reflection on AWS's early struggles offers more than just corporate nostalgia. At a time when AI investments are facing scrutiny and tech companies are under pressure to show immediate returns, his emphasis on controlling what you can control while persisting through uncertainty provides a roadmap for long-term thinking. The man who helped build one of tech's most profitable divisions is essentially telling employees and the broader industry that the next breakthrough might require the same uncomfortable patience that transformed a "nutty" cloud experiment into Amazon's crown jewel.
