Picture a warehouse the size of thirty football fields humming at 3 a.m., robots ferrying shelves to human pickers while, somewhere in the same building’s ceiling, fiber cables carry the data traffic of thousands of companies renting Amazon’s computers by the minute. That dual image, the cardboard box and the cloud server, is the whole Amazon story. For most of the last decade, Wall Street paid a sky-high price for the box and got the server almost for free. Then the server started printing money, and everything changed.
Amazon is scheduled to report its second-quarter 2026 results on July 30, 2026. The stock closed Tuesday at $230.86, sitting meaningfully below its 52-week high and barely changed over the past year. The market is not panicking, but it is not celebrating either. That flat line hides a genuine disagreement between what the filed numbers say the business is becoming and what the price tag says investors are willing to pay for it. That disagreement is the story.
In one breath
Amazon’s filed results show a business that has roughly doubled its net margin in two years while its cloud division is growing faster than it has in years. The stock’s forward earnings multiple sits near the lowest decile of its own decade-long range, our data show; one reading of that multiple is a market expecting the improvement to be largely over. Those two readings are in tension: either the margin improvement continues and the multiple looks low, or the market’s skepticism proves correct and the improvement stalls, the filed data alone cannot resolve which. The enormous capital spending plan for 2026 is the reason the market hasn’t simply closed the gap.
How the server became the story
Three years ago, AWS growth had slowed to a pace that made investors nervous. In the first quarter of 2024, AWS grew 17% year over year, solid but unremarkable for a business that had once compounded at twice that rate. The retail operation was squeezing out cost, but the cloud was the engine, and the engine felt like it was cooling.
What followed was a reacceleration that changed the valuation conversation entirely. By the first quarter of 2026, AWS was growing 28% year over year, its fastest pace in fifteen quarters. CEO Andy Jassy said in April 2026 that “demand currently exceeds capacity” for AWS, a statement that, if it holds, means the constraint on growth is concrete and fiber, not customers. The AWS contracted backlog entering 2026 stood at $244 billion, up 40% from a year earlier, roughly four times what the division earns in a year. That backlog is the closest thing to a receipt the business has for future revenue.
For a saver trying to understand why any of this matters: AWS is the part of Amazon that rents computing power to other businesses the way a utility rents electricity. When a company builds its AI tools on Amazon’s servers, it pays by the hour, every hour, for years. The backlog figure is the stack of signed contracts for those future hours. A backlog growing faster than revenue means the pipeline is filling faster than the tap is flowing, which is a good problem to have, provided the pipes get built in time.
The capex paradox: spending a year’s revenue in a single year
Here is where the math gets uncomfortable. Amazon has guided for roughly $200 billion in capital expenditures for 2026, a figure The Motley Fool predicted on July 27 that CEO Jassy may raise on the Q2 2026 earnings call. To put that in human terms: it is more than Amazon earned in total revenue in any single year before 2021. The company is proposing to spend, in one calendar year on construction and equipment, an amount that would have bought the entire business twice over a decade ago.
In the first quarter of 2026 alone, capital expenditures reached $44.2 billion, up 76% from the same quarter a year earlier. The consequence is visible in the cash flow statement: trailing twelve-month free cash flow, the money left after all that building, has collapsed from a healthy level a year prior to nearly nothing. Free cash flow is what a business has left to return to shareholders or pay down debt after it has paid for everything it needs to keep running and growing. A number that falls that far in a year is not a rounding error; it is a strategic choice, and the market is deciding whether to trust it.
As IG Bank Switzerland noted on July 28, analysts are framing this as a “CapEx Paradox”: the spending that crushes near-term cash flow is the same spending meant to lock in the pricing power and capacity that justifies the long-term multiple. The bet is that today’s concrete becomes tomorrow’s margin. That bet has a historical precedent inside Amazon itself: the company spent years losing money on fulfillment infrastructure that eventually became the logistics moat that made one-day delivery a standard rather than a luxury.
What the price tag actually says
The math, in our framework, is what the filed fundamentals justify. The mood is what the market is currently paying. Right now, the mood is cautious in a specific, quantifiable way.
At $230.86, our data show Amazon’s trailing price-to-earnings ratio sits in the 11th percentile of its own decade-long range. In plain terms: the stock’s trailing P/E sits in the 11th percentile of its own decade-long range, meaning it has traded at a higher multiple 89% of the time over that period, a data point readers can weigh alongside the capex and cash-flow considerations. Yet the price-to-sales ratio sits near its decade median, meaning the market is not discounting the revenue, just the profitability. That combination says the market believes revenue will keep coming but is skeptical that the margin improvement of the last two years will continue at the same pace.
Analyst consensus, as compiled by Barchart as of July 2026, projects meaningful earnings growth through FY2027, including expectations that AWS growth accelerates further in both 2026 and 2027. If AWS accelerates through 2026 and 2027 in line with those published projections, and operating margin holds at or above the level reported in the FY2025 10-K, applying the same trailing multiple to those higher consensus earnings figures produces a range of arithmetic outcomes materially above today’s close, a scenario readers can compare against the bear-case arithmetic that follows. If either assumption slips, the identical arithmetic runs in reverse. The capex risk outlined below is the variable that determines which branch the arithmetic follows. If capex stays elevated and free cash flow stays near zero, the earnings number exists on paper but the cash does not, and multiples applied to paper earnings have a way of compressing.
Seeking Alpha, in a July 28 piece, argued an intrinsic value of $155 per share, citing the capex burden as the central concern. That view sits well outside the sell-side consensus, where MarketBeat’s aggregation of analyst valuation estimates, as of July 2026, shows a wide dispersion around a central tendency that sits above today’s close, a spread that itself reflects the unresolved capex debate rather than a settled view. The gap between the bear case and the bull case is not noise; it is a genuine disagreement about whether the capex converts to durable margin or simply inflates the asset base without proportionate return.
Reading the numbers
Net margin: 10.8% in FY2025 (filed in the FY2025 10-K). Net margin is what a business keeps from each dollar of sales after every cost, tax, and interest payment. Two years earlier it was 5.3%. Think of a shop that kept $5.30 from every $100 of sales and now keeps $10.80: same shop, same customers, twice the take-home. That improvement is the core of the bull case.
AWS Q1 2026 operating margin: 37.7% (filed in the Q1 2026 10-Q). Operating margin is what a division keeps after its own direct costs, before corporate overhead and taxes. A margin near 38% means that for every $100 of cloud revenue, AWS keeps $38 as operating profit. A grocery chain earning 2% would need to sell fifty times more groceries to match that dollar. The question for the Q2 2026 report is whether Q2’s margin, expected near 34% per analyst consensus, signals a structural dip from capex pressure or a temporary one.
Q1 2026 capex: $44.2 billion in a single quarter (filed in the Q1 2026 10-Q). A household spending $200 a month on utilities would need to spend the equivalent of $176,000 in a single month to match the proportional jump Amazon made in capital spending versus its prior pace. That is the scale of the commitment. If the next two quarters match Q1, the full-year total clears $175 billion before any raise Jassy might announce on the July 30 call.
P/E at the 11th percentile of its own decade, our data show. A percentile here works like a class rank: if Amazon’s P/E has been higher than today’s reading 89% of the time over the past ten years, on that single measure the current multiple sits in the low end of its own decade-long range, a data point that can be read as relative compression or as a reflection of changed earnings quality, depending on the analytical framework applied. The catch is that the prior decade included years when Amazon earned almost nothing, so the median was inflated by optimism about future profits. Today’s lower multiple may simply reflect a market that has learned to wait for the cash, not just the earnings line.
For the standing yardsticks on Amazon: the price tag, the filed record, and the four gauges, refreshed with each edition, see the BullScope Evidence Sheet: Amazon.
Sources
- Amazon 10-Q, Q1 2026 (SEC EDGAR)
- Amazon 10-K, FY2025 (SEC EDGAR)
- Amazon 10-K, FY2023 (SEC EDGAR)
- IG Bank Switzerland, Amazon earnings preview, July 28, 2026
- The Motley Fool, capex raise prediction, July 27, 2026
- Investing.com, AWS backlog and margin outlook, July 2026
- Barchart, Amazon analyst estimates, July 2026
- MarketBeat, Amazon analyst valuation estimate aggregation, July 2026
- Seeking Alpha, intrinsic value analysis, July 28, 2026
- GeekWire, Amazon Q1 2024 results, 2024
