Amazon Web Services has been here before. Heavy spending on data centers, a surge in demand tied to a new technology wave, and then years of expanding profits as customers fill the capacity. The pattern played out during the enterprise cloud migration of the mid-2010s and again during the pandemic-era digital rush. Now, with artificial intelligence reshaping how companies use computing power, AWS appears to be running the same playbook, only at a scale that dwarfs every previous cycle.
The financial foundation for that comparison is laid out in Amazon’s 2024 annual report filed with the SEC. AWS posted $107.6 billion in net sales for the year, a 19% increase that marked its strongest annual growth rate since 2021. Operating income jumped 77% to $39.8 billion, pushing segment margins above 37%. Those are the kinds of numbers that make the AI investment thesis look less like speculation and more like a bet with historical precedent.
How the acceleration built through 2024
The growth did not arrive all at once. AWS’s revenue run rate climbed past $110 billion on an annualized basis by September 2024, according to the company’s third-quarter 10-Q filing. Management pointed to rising demand for compute capacity tied to generative AI workloads, layered on top of continued strength in traditional cloud services like storage and databases.
But the margin story requires a closer look. Amazon disclosed in its 10-K that it extended the estimated useful life of servers and networking equipment during 2024, reducing depreciation expense by roughly $2.8 billion across the entire company. Because AWS accounts for the majority of Amazon’s infrastructure assets, a substantial portion of that benefit landed on the cloud segment’s income statement. Strip out the accounting change, and margins still improved, but not by as much as the headline number suggests. That distinction matters for anyone projecting whether 37%-plus margins are sustainable or partly a one-time lift.
Even with that caveat, the underlying momentum was real. Revenue growth reaccelerated after a stretch in 2022 and early 2023 when enterprise customers pulled back spending to optimize existing cloud contracts. The workload mix shifted toward higher-value AI services, though Amazon does not disclose AI-specific revenue within AWS. The annual filing attributed the segment’s performance in part to “increasing demand for technology infrastructure driven by the expansion of generative AI.”
The pattern that keeps repeating
AWS launched its first commercial services in 2006. Over the roughly two decades since, the business has moved through at least two distinct investment-and-harvest cycles.
The first ran from about 2014 to 2016. Amazon poured billions into data center construction to meet a wave of enterprise workloads migrating off on-premises servers. Margins compressed during the buildout phase, then climbed as utilization rates rose and customers locked into long-term commitments. The second cycle arrived in 2020 and 2021, when pandemic-driven digital adoption filled newly built capacity faster than Amazon had projected.
The current AI cycle follows the same arc, but the numbers are far larger. Amazon’s total capital expenditures reached $83 billion in 2024, with the bulk directed at AWS infrastructure. During the company’s February 2025 earnings call, CEO Andy Jassy told analysts, “We expect to invest approximately $100 billion in capex in 2025, with the majority focused on AI infrastructure to meet the growing demand we’re seeing from customers.” For context, Amazon’s total capex in 2016, near the peak of the first cycle, was roughly $7 billion.
The logic behind the spending is straightforward: every prior wave of heavy investment produced durable margin expansion once customers adopted the new capacity. The open question is whether AI demand will follow the same sticky adoption curve that enterprise cloud migration did, or whether it will prove more volatile as companies experiment with generative AI use cases that have not yet settled into production-grade workflows.
Enterprise AI adoption on AWS remains hard to quantify
Amazon’s SEC filings are detailed, but they leave critical questions unanswered. AWS is reported as a single operating segment, which means there is no public breakdown separating revenue from AI-native services like Bedrock (Amazon’s managed foundation model platform), SageMaker (its machine learning toolkit), or custom silicon such as Trainium and Inferentia chips from legacy compute, storage, and database products. Without that split, it is impossible to know whether AI is the primary engine of AWS’s reacceleration or one tailwind among several.
What is visible from public disclosures and Amazon’s own statements is that large organizations have been building AI workloads on AWS infrastructure. During the February 2025 earnings call, Jassy highlighted that AWS’s AI business had reached a “multi-billion-dollar annual revenue run rate” and described demand as growing at a “triple-digit year-over-year percentage.” He cited customers across financial services, healthcare, and media adopting Bedrock and custom training on AWS infrastructure, though Amazon did not name specific accounts or disclose contract values in its filings. That framing suggests broad enterprise uptake, but without named customer wins or deal sizes in the public record, the depth of adoption remains difficult to independently verify.
Capital allocation is similarly opaque. Amazon reports total property and equipment purchases but does not separate spending on GPU clusters and AI-optimized facilities from routine capacity expansion. That makes it hard to calculate the return on AI-specific investment or to gauge how much of the near-term margin pressure comes from AI buildouts versus general infrastructure growth.
Competitive positioning across multiple quarters
The competitive landscape adds another dimension. Microsoft reported that its Azure cloud platform grew 33% in the December 2024 quarter, with management attributing 13 percentage points of that growth directly to AI services during the January 2025 earnings call. That data point is now more than a year old relative to spring 2026, and subsequent Microsoft earnings reports have shown Azure’s AI contribution continuing to expand, though the precise figures for more recent quarters sit outside the scope of the 2024 filings analyzed here.
Alphabet’s Google Cloud crossed $40 billion in annualized revenue during the same late-2024 period. Both companies have sustained aggressive AI infrastructure spending through 2025 and into 2026, competing for the same enterprise contracts that AWS is pursuing. The broader trend across multiple quarters is consistent: all three hyperscalers have reported AI as a growing share of cloud revenue, but none has disclosed enough detail to allow precise market-share calculations for AI-specific workloads.
AWS remains the largest cloud provider by total revenue. Microsoft’s deep partnership with OpenAI and Google’s Gemini model family give each rival differentiated positioning in AI-specific workloads where competitive lines are still being drawn. Amazon’s counter is its Anthropic partnership and its custom chip strategy, but the relative commercial traction of each approach is not yet clear from public filings alone.
What has and has not changed since the 2024 filing
Between the publication of Amazon’s 2024 annual report in early February 2025 and spring 2026, several quarters of additional results have accumulated. Amazon reported full-year 2025 results and first-quarter 2026 earnings during this window, and the broad trajectory has held: AWS revenue has continued to grow, and capital spending has followed through on the commitments Jassy outlined. Amazon’s partnership with Anthropic, which has involved a total investment commitment of up to $4 billion and deep integration of Anthropic’s Claude models into the Bedrock platform, has expanded in scope, though the financial impact of that relationship is still not broken out in segment results.
What has not arrived is the kind of granular AI revenue disclosure that would let investors definitively measure how much of AWS’s strength traces to artificial intelligence versus the broader maturation of cloud computing. The depreciation question also remains live. Useful life extensions provided a measurable tailwind to 2024 margins, and any reversal or further adjustment in future filings could shift the profitability picture materially. Investors parsing AWS results should watch for updated depreciation assumptions in subsequent 10-K filings, where accounting choices can move operating income by billions of dollars.
What two decades of AWS investment cycles suggest about the AI era
The most useful framework for evaluating AWS right now is not any single quarter but the pattern across its full history. Every major technology shift that pushed enterprises toward cloud infrastructure, from mobile computing to big data analytics to containerized applications, produced a cycle of heavy AWS investment followed by years of expanding profitability. AI appears to be following that template.
The important caveats are real: the capital commitments are an order of magnitude larger than anything in prior cycles, the competitive field is more crowded, and the durability of enterprise AI spending has not been tested through a full economic cycle. Amazon’s audited filings confirm the financial foundation, with accelerating revenue, expanding margins even after adjusting for depreciation changes, and a willingness to invest at a pace no competitor has matched. What those filings cannot confirm is whether the AI boom will prove as profitable per dollar invested as the cloud adoption waves that came before it. That is the question that will define AWS’s next chapter, and answering it will require both time and the kind of disclosure Amazon has so far been reluctant to provide.
David M. Keller is a finance writer based in Columbus, Ohio, covering personal finance and consumer-focused economic topics. He earned his degree in journalism from Ohio University and began his career reporting on local business and economic trends for a regional media outlet. Since then, he has contributed to a variety of online publications, focusing on clear, practical coverage of topics such as cost of living, debt, and everyday financial decision-making.
