Is Amazon Stock Overvalued?
Amazon (AMZN) stock is trading at $241, near its all-time high. Even on a relative basis, the stock’s price-to-sales multiple has climbed to a near three-year high of 3.8x, up from levels of about 2x in 2023. (Amazon stock valuation)
Sure, the multiple has been higher, at levels of over 5.4x at the height of the post-pandemic expansion, before rate increases reset valuations across growth equities. The current multiple is not extreme by that standard. But it is a prompt to ask what has changed fundamentally to justify a premium toward the top of the recent range.
The market is paying more for Amazon’s top line because the quality of its revenue has fundamentally changed. The company is no longer just a retailer with a cloud side hustle; it has evolved into a high-margin AI utility.
Margins Tell The Story
The answer lies in margin structure. In 2022, Amazon’s consolidated operating margin was near 2%. By 2025, it had expanded to approximately 11%, driven by faster growth in Amazon Web Services and the advertising segment, both of which carry significantly higher margins than the retail business. EPS moved from a loss of $0.27 in 2022 to $7.29 in 2025.
AWS And The Silicon Advantage
AWS is the clearest driver. As generative AI remains an intensely compute-heavy frontier, Amazon’s position as the world’s largest infrastructure provider makes it the natural beneficiary of this transition. Revenue growth reaccelerated to 24% in the most recent quarter, reaching a $142 billion run rate.
Amazon’s custom silicon business, which includes Trainium and Graviton chips, hit a $20 billion annualized run rate. Proprietary silicon reduces dependence on third-party GPU suppliers while allowing it to run AI inference workloads at lower cost, giving enterprise customers a more economical path to deploying models at scale.
Inference opens the door for many players to challenge AI titan Nvidia (NVDA) and Broadcom (AVGO) ASICs are seen as the primary challenger. See How Broadcom Ends Nvidia’s GPU Monopoly
CEO Andy Jassy noted in his shareholder letter that if this division were treated as a standalone hardware supplier, its run rate would approach $50 billion.
The Capital Cycle Repriced
Capital expenditures reached $131.8 billion in 2025, with management guiding toward approximately $200 billion in 2026. That level of spending compressed free cash flow sharply last year and weighed on sentiment.
The market’s interpretation has since shifted: investors are now treating the capital cycle as evidence of demand visibility rather than financial risk, particularly given Jassy’s disclosure that a substantial portion of the new capacity is already committed by customers. Committed revenue before infrastructure is built out changes the risk profile of the spending materially.
The Bottom Line
The thesis now hinges entirely on execution. Can Amazon translate this capital investment into visible operating leverage by 2027? If cloud utilization rates remain high and Trainium silicon delivers the margin improvement the company is projecting, the current valuation is justified. If capacity sits idle, depreciation costs will pressure earnings, and the stock will face headwinds.
At 3.8x sales and 35x trailing earnings, Amazon is not expensively priced relative to its own history. But it is priced for continued execution, with little room for the capital cycle to disappoint.
For investors seeking exposure to Amazon’s AI infrastructure thesis without taking a concentrated single-stock position, the broader question is how to capture that upside while managing execution risk.
The Trefis HQ strategy offers one such approach, having outperformed its blended benchmark (S&P 500, S&P MidCap, and Russell 2000) with over 105% returns since inception. Why did HQ outperform? See HQ performance metrics with five reasons why – for the full story.