TradingKey - From a straightforward price-to-earnings perspective, Amazon (AMZN) does not represent a cheap stock at nearly 35x trailing earnings and about 31x forward earnings.
This metric does not consider how invested Amazon has been in the most aggressive period of its history in terms of the introduction of artificial intelligence, as evidenced by its Fourth Quarter results, which illustrate how the investments will eventually translate into significant scale.
Consolidated net sales grew 14% on a year-over-year basis to $213.4 billion in the Fourth Quarter of 2025. Amazon Web Services revenues increased by 24% to $35.6 billion (compared to 20% in the prior quarter).
The growth of Amazon's subscriptions and advertising also continues to rapidly expand alongside Amazon's cloud business, yet these two high-margin revenue-generating businesses often do not receive the same level of focus as the company's cloud business. Despite AWS being a smaller percentage of total revenue, the cloud business continues to provide significant margin dollars and profits to the overall business - $12.5 billion in operating income in the Fourth Quarter, which represented approximately half of total operating income for Amazon in that quarter.
Over the course of 2025, AWS represented approximately 18% of total sales but generated about 57% of total operating income, which further emphasizes how crucial the company's cloud business continues to be to its investment thesis.
This year, investors have been focusing their attention on three main trends: First, the demand for AI (artificial intelligence) technologies is increasing rapidly; Second, AWS (Amazon Web Services) is rapidly expanding its network capacity; Lastly, many AI model developers (like Anthropic) are establishing stronger relationships with AWS as they rapidly scale.
KeyBanc Analyst Justin Patterson believes that the combination of these three major trends has resulted in a net benefit to usage of AWS cloud services (i.e., there has been an increase in demand) — as illustrated by the rapid growth of Anthropic.
According to KeyBanc data, Anthropic had $9 billion of recurring revenues at December 31, 2025, and then increased to $30 billion at the beginning of April 2026. Approximately 60% of Anthropic's estimated spending on its cloud services is through the AWS cloud (i.e., the AWS cloud is pulling approximately 60% of all cloud service spending at Anthropic). Thus, the rate at which Anthropic has continued to release products provides evidence that customers are continuing to use their cloud providers more frequently.
In March 2026, Anthropic released Claude Opus (4.7) — the largest and most advanced reasoning model ever produced — along with Claude Mythos (a hyper-agentic application that is believed to be classified for national security reasons). The significance of these releases is that they are expected to lead to an increase in AWS usage over the next few months.
There are also broader signs concerning the expansion of AI infrastructure. For example, TSM's recent earnings report had strong results; there is no observable downward demand trend in chips, hardware, or all AI-related software (according to Dan Ives at Wedbush), which he believes gives investors great "green light" signals for continued ownership of core tech leaders until Q1 earnings appear. Consequently, these same elements have resulted in Amazon's stock increasing 16% from such news.
Increasing AI capabilities at Amazon won't just involve renting now but also making Trainium chips available for third parties, representing another avenue for growth, as made clear in Jassy's annual August 2023 letter to shareholders.
With chip revenue through AWS already having exceeded $20 billion and experiencing triple-digit year-over-year growth, there is clear evidence that silicon and infrastructure can provide material benefit.
On the demand side, Amazon also has close ties, both financially and commercially, with many of the major model builders. For example, since the end of 2023, Amazon has invested $8 billion into Anthropic and at year's end had $45.8 billion in convertible notes and $14.8 billion in non-voting preferred shares of Anthropic for a total valuation of $60.6 billion.
According to Yahoo Finance, Anthropic was given a valuation of $380 billion after receiving $30 billion in funding in February (making it the third-largest valued private business), and it's reported that investors have expressed interest in an $800 billion valuation. Given how often private values change, the fact that Anthropic usage is in a lot of ways linked to AWS as well as Amazon’s stake in Anthropic, gives Amazon many different ways to benefit from expected ongoing demand for AI.
Amazon’s stock price expectations are to balance its massive scale with its rapid growth trends that are difficult to predict in the coming 12 months or so. A more reasonable rate for Amazon stock over this time period would be in the mid-teens total return range, but for this to happen, AWS needs to continue or accelerate its growth from current levels (this would validate the long-term ROI from current capital spending).
At $252/share now, assuming 12% annual compounding returns, you would expect Amazon’s stock to be $282.24/share in approx 12 months and $444.58/share in approx 5 years.
The above estimates should not be viewed as exact values; they are meant as estimates given the fact that they also assume both capacity additions will continue to grow to meet ongoing strong demand levels, AI workloads will continue to be highly robust and that operating leverage relative to depreciation will improve as normal total depreciation rates occur over the next several years.
In Q1 ’26, analysts expect high growth rates of AWS continuing at approx 30% per annum through Q1 ’26 based on many different factors, including the fact that AWS has a good outlook for further AWS Accelerated Growth in ’26 if supply chain bottlenecks ease and AI use continues to increase.
Significant capital expenditure may cause reported earnings to be under pressure for several years as depreciation accumulates, and there is no assurance that every dollar of investment will generate a return as planned. Supply chain constraints can inflict damage either by delaying delivery or by enabling competitors to close the operational hole. The stock trades at a non-bargain multiple of earnings and therefore, a change in sentiment can result in an outsized change in the multiple.
A high degree of concentration across multiple avenues of AI demand creates a cyclicality of sorts; strong indicators from end customers, such as Taiwan Semiconductor, could become irrelevant due to changing macro-economic conditions. Strategic investments in privately-held companies are subject to mark-to-market fluctuations from time to time, and some of Amazon’s present-day partnerships that are driving growth through AWS may become repriced by their customer partners over time.
While none of these risks are brand new to Amazon, they will potentially have a greater impact because of elevated expectations for both growth and capital intensity.
An investor looking to buy into Amazon today should evaluate which metric depicts the story they are purchasing correctly.
Metrics based on P/E would indicate shares being overvalued at peak buildout; however, if one looks at what the operating cash flow multiple indicates, you will see a much more reasonable representation of future growth with respect to its rapidly increasing volume of AWS cloud services, chips and infrastructure and an increase in subscriptions and advertising as well as a broad range of rapidly expanding AI customer prospects.
The key positive attribute supporting the idea of owning the stock is that AWS will continue to have the momentum that it has experienced so far but at higher growth rates due to management producing attractive returns on their investment to date with approximately $200 billion of capital expenditures (to be completed within this year) and longer-term depreciation rates decreasing as the current build transitions into a much more efficient operational environment and with margins expanding in the process of achieving maximum levels of operating capacity utilization.
The market appears to offer a possible opportunity for long-term investors, with expected compounded returns of mid-teens over the long-term, assuming no perfection whatsoever.
New information has become available through quarterly reporting, in the case of the most recent report from April 29, 2026, over several weeks; as a result, there were some minor misalignments between performance based on the new data and performance based on previous reports.
The risk-reward profile for a long-term investor who can withstand the noise associated with quarterly results in terms of generating predictable long-term cash flow is still very attractive, especially taking into account that the two most important factors driving growth are continued execution of capital investments and continued growth of demand for AI in the future.