Amazon reported second-quarter 2026 adjusted earnings of $1.88 per share against a consensus of $1.83, on quarterly net sales of $200.6 billion — beating the revenue consensus by about 1.8%.
Two hundred billion dollars in ninety days. It is worth sitting with that number for a second before analyzing it, because the scale is genuinely difficult to hold in your head: that is roughly $2.2 billion of revenue per day, every day, for a quarter.
The quarter
| Metric | Result | Expected |
|---|---|---|
| Net sales | $200.6B | beat by ~1.8% |
| Adjusted EPS | $1.88 | $1.83 |
| Implied daily revenue | ~$2.2B | — |
What a beat means at this size
A 1.8% revenue beat sounds modest until you convert it: roughly $3.5 billion of revenue above what the Street modeled, in a single quarter, from a base that everyone has been forecasting for two decades.
The thing that has changed about Amazon over the last several years — and that the market has been slow to internalize — is that the earnings line is no longer driven by retail. It is driven by AWS and by advertising, two businesses with completely different margin structures bolted onto a low-margin logistics operation. When the earnings beat and the revenue beat move together, it usually means the high-margin businesses carried the quarter rather than the retail volume.
That is why the earnings surprise is worth more than the revenue surprise here. Retail revenue can be bought with promotions. AWS and ad revenue cannot.
The part I'd be careful about
The same thing I flagged on Microsoft applies here with equal force: the capital expenditure required to serve AI demand is enormous, it is happening simultaneously across every hyperscaler, and it converts into depreciation that sits on the income statement for years.
Amazon has an additional structural exposure the others don't — the retail business is consumer-facing and rate-sensitive in a way cloud infrastructure isn't. Into a period where consumer signals have been mixed and a Fed decision is live, the retail half of this business is levered to a variable that has nothing to do with AI.
There is also a persistent margin question. Amazon can generate enormous revenue at will; the entire investment case for the last decade has been about whether it converts that revenue at rates comparable to its peers. A $1.88 print against $1.83 is a good answer for one quarter, not a settled one.
My take
This is a solid quarter from a business whose main risk has never been demand. I don't find $200.6 billion in revenue surprising so much as clarifying — it establishes that the top line is not the constraint and never was.
What I watch on Amazon is the gap between AWS growth and total capex. Same discipline as Microsoft: while cloud revenue growth runs at or above the capex ramp, the spending is investment. When capex accelerates into decelerating cloud growth, it becomes a fixed-cost problem in a business with thin retail margins underneath it.
For now the print says the machine is working on both ends — the high-margin businesses beat and the scale keeps compounding. I'd own it for the second-order businesses, not for the retail volume that generates the headline.
Bottom line: $200.6B of revenue and a beat on both lines. The retail number is the headline; AWS and advertising are the earnings.
This is analysis, not investment advice.