Published May 2026
For the better part of two years, the knock on Amazon Web Services was that the cloud giant’s best growth days were behind it. The first quarter of 2026 just dismantled that narrative.
AWS generated $29.3 billion in revenue during the first three months of the year, a 28% increase from the same period in 2025 and the cloud division’s fastest expansion in 15 quarters. On the earnings call, CEO Andy Jassy disclosed that AI services within AWS have reached an annualized revenue run rate exceeding $15 billion, a figure that, on its own, would rival the total annual revenue of major enterprise software companies like ServiceNow.
The results, disclosed in Amazon’s quarterly SEC filing, represent a sharp reversal from the slowdown that dogged AWS through much of 2023 and into 2024. During that stretch, enterprises pulled back on cloud spending and squeezed more value out of existing contracts. The last time AWS posted comparable growth was the second quarter of 2022, when pandemic-driven digital migration was still fueling expansion across the industry.
AI spending moves from experiment to budget line
The $15 billion annualized AI run rate is a management-defined metric, not an audited line item, but the scale is difficult to wave away. It signals that corporate customers have moved well past testing AI tools on AWS. They are now spending at volumes large enough to visibly shift the revenue trajectory of a division pulling in roughly $115 billion a year.
That spending falls into two broad categories, and the mix between them matters more than the headline number. Training large AI models requires enormous bursts of compute power, often sustained over weeks or months, generating intense but episodic demand. Inference, where trained models serve real-time predictions to end users, produces steadier, more recurring revenue. Amazon has not disclosed the split. A run rate tilted toward training could prove volatile quarter to quarter; one anchored by inference would suggest AWS is building a durable new revenue stream atop its existing cloud business.
Amazon has been investing to capture both types of demand. Its custom Trainium and Inferentia chips are designed to undercut Nvidia GPU pricing for specific AI workloads, and its Bedrock platform gives developers managed access to foundation models from Anthropic, Meta, and others. During the earnings call, Jassy pointed to growing adoption of these services among startups and large enterprises alike, though he stopped short of naming customers or providing granular revenue breakdowns.
The capital question behind the growth
Sustaining 28% growth in a business this large does not come cheap. Amazon previously outlined plans to invest more than $100 billion in capital expenditures during 2025, with a significant share directed toward data centers, networking infrastructure, and AI chip capacity. Early indications suggest that spending pace has carried into 2026, though the Q1 filing does not isolate AI-specific capital outlays.
That creates a tension investors are watching closely. The top-line acceleration validates the thesis that AI is reigniting cloud spending, but the capital required to sustain it could weigh on free cash flow and returns on invested capital in the near term. And because Amazon does not break out AI revenue as a separate audited segment, analysts building valuation models around the $15 billion run rate are working with management estimates rather than GAAP figures.
Sid Nag, a cloud infrastructure analyst at Gartner, noted in a May 2026 research briefing that AWS’s AI run rate “validates what we have been hearing from enterprise clients: AI workloads are no longer discretionary pilots but line items in annual cloud budgets.” Nag cautioned, however, that the lack of audited segmentation makes it difficult to compare AI revenue claims across hyperscalers on an apples-to-apples basis.
Amazon’s broader first-quarter results offered some cushion. The company reported increased profits and total net sales, with AWS continuing to generate the majority of operating income even as the retail and advertising segments grew. That profitability buffer gives Amazon more room than most competitors to fund massive infrastructure buildouts without immediately squeezing margins.
Where AWS stands against Azure and Google Cloud
The 15-quarter growth record positions AWS as reaccelerating, but the competitive picture is still taking shape. Microsoft and Google had not yet reported their comparable quarterly results at the time of Amazon’s disclosure. Microsoft had previously disclosed an AI annualized revenue run rate of $13 billion in early 2025, meaning Amazon’s $15 billion figure, if the metrics are roughly comparable, suggests AWS has at minimum kept pace and possibly pulled ahead on AI-specific cloud revenue.
All three hyperscalers are racing to expand AI capacity. Microsoft has poured tens of billions into its partnership with OpenAI and Azure’s AI infrastructure. Google has leaned into its Tensor Processing Units and Gemini model family to differentiate Google Cloud. For enterprise buyers, the aggressive buildouts across all three providers could translate into pricing pressure and better contract terms, particularly for companies willing to commit to multi-year deals.
What the quarter does and does not prove
A single quarter of 28% growth does not guarantee a sustained trend. Enterprise AI adoption remains early enough that large one-time migrations or model-training projects could inflate any given period’s results. Chip supply constraints, pricing competition, and the pace at which companies move AI workloads from pilot to production will all shape whether AWS can maintain this trajectory through the rest of 2026.
The durability question extends to margins. Revenue from higher-level managed services, such as hosted model inference or proprietary AI platforms, typically carries better margins than raw compute and storage. How much of the AI run rate comes from premium services versus commodity infrastructure will determine whether the AI surge ultimately lifts AWS profitability or simply adds revenue at thinner margins.
AWS makes its case as AI’s infrastructure backbone
Strip away the caveats and the Q1 numbers establish something concrete: AI has crossed the threshold from corporate talking point to material revenue driver inside Amazon’s most important business unit. The $15 billion run rate, the 28% growth clip, and the capital commitments behind them all point in the same direction. The cloud market’s next phase will be defined not by generic workload migration but by specialized, compute-intensive AI services. Amazon just posted the strongest evidence yet that it intends to set the terms for that phase, not just participate in it.
