For most of the last decade, I looked at Amazon (AMZN +0.39%) and thought, "Amazing business, but too messy for me as a shareholder."
I saw razor-thin retail margins, huge capital spending, and a company that seemed to reinvent itself every other year.
Now I find myself changing my mind, not because the stock has gone up, but because the underlying business looks very different than the one I kept passing on.
Image source: Getty Images.
The profit engine I underestimated
For years, I treated Amazon as just a retailer. That was my first mistake. Today, its own filings make it clear that the profit center is Amazon Web Services (AWS), not cardboard boxes. In 2025, AWS generated roughly mid-double-digit billions in operating income, far more than the retail segments, and did so with margins that look more like a software company than a store.
What changed my thinking was seeing that AWS is not a side hustle layered on top of e-commerce. It is the backbone of a huge part of the internet, with long contracts, deep integration, and economics that can fund a lot of experimentation elsewhere.
When I used to worry about Amazon's spending, I did not fully appreciate that a high-margin engine was quietly paying the bills.
NASDAQ: AMZN
Key Data Points
A quiet advertising giant
The other blind spot for me was advertising. I always thought of ads as something that mattered for Alphabet and Meta, not Amazon. Then I started reading Amazon's own numbers and language around "advertising services" and "retail media." Earlier this year, Amazon disclosed more than $60 billion in annual ad revenue, growing at roughly 20% year over year and outpacing the core business.
That is not just banner clutter. It is brands that are paying for a front-row seat at the moment of purchase. Amazon sits where intent is strongest. The more I thought about that, the more I realized this might be one of the most durable profit streams for the company, built on shopper data and closed-loop measurement that is hard to replicate.
The logistics moat I used to see as a cost problem
My last hang-up was the warehouses, planes, and vans. I saw a cost monster. Over time, Amazon's history of its fulfillment and delivery network started to read differently. This is not just a way to move packages. It is an infrastructure grid that gets products closer to customers than most rivals can, and now supports third-party sellers, same-day delivery, groceries, and more.
There are still real risks: massive capital spending, regulatory scrutiny, and a business culture that demands constant reinvention. But after 10 years of standing on the sidelines, I now see a company whose profit mix has shifted toward AWS and advertising, with a logistics network that looks more like a moat than a burden. For me, that combination finally justifies owning Amazon as a long-term core holding, sized reasonably, instead of treating it as a great business I never quite trust enough to buy.
