Amazon’s twin engine strategy—combining massive internal AI adoption with aggressive infrastructure investment—is proving highly effective in driving growth and maintaining dominance in the cloud and AI space. Like Google and Meta, Amazon has long used its own operations as a real-world testing ground for AI, refining systems that power everything from retail and logistics to advertising and cloud services. This internal use of AI, or “Customer Zero,” allows Amazon to better understand the needs of external customers, making AWS not just a cloud provider but a trusted partner in AI transformation. The recent layoffs of over 60,000 corporate employees are not solely about cost-cutting. While some of the reductions stem from post-pandemic over-hiring and ongoing workforce rebalancing, they also reflect a broader shift toward automation. As physical robots have already transformed Amazon’s fulfillment centers—growing from 200,000 in 2019 to over 1 million today—the next frontier is AI agents handling corporate tasks. The Rufus shopping assistant is just the beginning. If economic pressure mounts, AI will increasingly take on roles like procurement, scheduling, and even workforce optimization, replacing human labor in predictable, data-driven workflows. Despite these changes, Amazon continues to operate at scale with about 1.2 million warehouse workers—down from a peak of 1.3 million—showing that human labor remains essential for complex, adaptive tasks. But the trajectory is clear: AI will increasingly take over routine, repetitive, and data-intensive functions, both in warehouses and in back-office operations. A key enabler of this transformation is Amazon’s massive capital investment. In 2025, Amazon spent $134.73 billion on capital expenditures—an increase of 60.5% over 2024—with AWS accounting for around $115 billion of that. Of this, approximately $105 billion was allocated to AI infrastructure, meaning AI-related spending made up about 78% of AWS’s total capex. The remaining funds supported logistics and warehouse operations. This spending is translating into explosive growth in compute capacity. In the past year alone, AWS added 3.9 gigawatts of datacenter capacity, averaging just $29.5 billion per gigawatt—significantly less than the $45–60 billion per gigawatt spent by competitors like OpenAI and Anthropic. By the end of 2025, AWS had around 6 gigawatts of installed capacity, with plans to double it to 12 gigawatts by 2027. At an estimated $30 billion per gigawatt, this expansion will cost $180 billion—well within Amazon’s projected $200 billion 2026 capex budget. The efficiency gains are not just in scale but in performance. With advancements in AWS’s custom chips—Graviton (Arm-based CPUs) and Trainium (AI XPU)—and software optimizations, compute performance could increase 3 to 4 times over the next two years, even without proportional increases in spending. This means AWS is not only building more capacity but smarter, more cost-effective capacity. Revenue is following suit. AWS generated $128.73 billion in 2025, up 19.7%, with operating income growing 14.5% to $45.61 billion. In Q4 2025, revenue hit $35.58 billion, up 23.6%. A key driver is the surge in compute revenue, which now likely exceeds software revenue due to the high demand and premium pricing for AI hardware. The Trainium2 chip fleet, now at 1.4 million units, generated a $10 billion annualized run rate in 2025—$2.5 billion in revenue. Trainium3 is already ramping, with all capacity expected to be reserved by mid-2026. This dual strategy—building in-house AI chips while also offering third-party hardware—gives AWS a unique advantage. By undercutting Nvidia and AMD on cost and performance, AWS can deliver better margins and faster innovation. As more enterprises adopt AI, the demand for compute will grow exponentially, fueling a self-reinforcing cycle of investment, capacity, and revenue. Amazon’s strong advertising business—now a major profit center—also plays a crucial role. It helps subsidize the massive capital outlays needed for AI infrastructure, ensuring that AWS can continue investing at scale even as margins come under pressure. In short, Amazon’s twin engine strategy—internal AI adoption and aggressive infrastructure investment—is not just working—it’s setting the pace for the entire AI era. And as AWS continues to lead in compute, software, and innovation, its ability to deliver value to customers will only grow stronger.