Amazon's Real Story Sits in the Cloud, Not Retail
Most people think of Amazon as amazon.com, the place they shop. That retail side is the world leader, with margins around 40 to 42%, and it grew about plus 1% year over year. It is doing fine. It is not where the story is.
The story is cloud compute and AWS. Retail still brings in the bulk of Amazon's revenue, yet AWS drives nearly 60% of operating profit while making up only about 20% of revenue. That gap is why all eyes are on the cloud number.
Azure Is Closing the Gap
The big fear for Amazon investors, and the big hope for Microsoft investors, is that Amazon's huge lead in AWS is shrinking and Microsoft could pass it.
The data here tracks year-over-year change, not total demand. It charts the derivative, meaning how fast the change itself is moving. A straight positive line would mean steady growth over time. On that measure, Azure jumps off the chart at plus 67% year over year. AWS chugs along at plus 13%. Google is hanging in there.
One caveat on the plus 13% for AWS: this measures people going to those websites and looking to sign up. It tracks new customers. It does not capture existing customers expanding what they already use, which is hard for the tracking method to see, since business managers do not post on social media about spending more. So the real AWS increase is probably higher than 13%. Microsoft is doing very well on new customers. Amazon is doing well growing existing customers. Both are growing fast with existing customers.
This is likely a big catch-up quarter for Microsoft. AWS stays in the lead, but the lead is shrinking a lot, and that could sit poorly with investors.
The Growth Rate Numbers
Analysts expect about 31% growth for AWS. For comparison, Google Alphabet's cloud side grew 82%, and Azure grew 43%, both above expectations. If AWS comes in above 31%, that would not be a surprise. Amazon is already the biggest player in cloud, and the amount of compute everyone says they need keeps climbing.
Whether AWS prints 25% or 31% does not change the read. The direction is clear, the pie is growing so much that there is room for both Microsoft and AWS to do extremely well. The LikeFolio earnings score is bullish at plus 31. Retail is hanging in, AWS is growing.
CapEx Is Buying Future Revenue
Amazon plans roughly $200 billion in CapEx in 2026 to build things out. Amazon does what it always does with big expenses: it builds the thing itself and turns it into a business. That started with AWS, where the huge server bills pushed them to build their own. They did the same with delivery. Now they are doing it with GPUs and AI software through Trainium (their own chips) and Bedrock (their AI software service).
This is a triple win for Amazon. A customer on AWS pays for the services, pays for the software side with Bedrock, and pays for the hardware side with Trainium. The customer forks over money to Amazon and its own business runs much better because of the AI.
When Meta reported, it beat on revenue and raised CapEx by about $10 billion at the bottom end, and the stock sold off. If Amazon also raises CapEx, the same negative reaction could follow. I read that as an opportunity. Increased spend here is an investment. These companies will not go bankrupt from spending too much; the risk is not there because so much revenue is coming in. They are buying future revenue with current expenses, and that is a fantastic model. Any sell-off on higher CapEx is a wrong read and a chance to buy.
AI compute matters more than many investors think, over the next year and beyond. The amount of AI used to build and analyze data sets is growing week over week, not just quarter over quarter. At large companies the growth is likely enormous. People are going all in because the results are so strong. A print above 31% on AWS growth should help relieve fears about overspending as the returns start showing.
Not a Zero-Sum Game
A common view coming into the year held that Amazon was the Magnificent Seven member with the most upside and investor appetite, and a lot of that rested on AWS. Amazon has been one of the laggards among the seven this year. Looking at the chart, Microsoft appears to be lapping Amazon.
The worry that if Microsoft wins, Amazon loses, is a limited and incorrect view. Compute is not a zero-sum game. The compute pie will grow exponentially over the next several years. Even if Microsoft takes the number one spot, Amazon can stay extremely profitable and keep growing. As an Amazon investor, I hope they do not lose the top spot, but I want everyone growing, because broad growth confirms the pie is expanding. You would not want to see Google and Microsoft fall off while only Amazon wins, because that would make you question whether the pie is growing at all.
One more footnote: Amazon owns roughly 15 to 20% of Anthropic, probably the best AI model out there right now. If Anthropic goes public at maybe a trillion dollars or more, that is $150 to $200 billion in paper value flowing to Amazon.
Prime Day was successful, up about 9% year over year, but it now reads almost like a footnote for the company.
In the short term, investors could misread the CapEx and growth signals. If CapEx rises and Microsoft shows it is outpacing AWS, and the stock sells off, that is an opportunity. A year from now, AWS's size will look far larger than it does today. The growth is coming.
