AWS Smashes Q2 Estimates as AI Demand Fuels 37% Revenue Surge

AWS Posts 37% Revenue Jump as AI Bets Start to Pay Off

Amazon Web Services delivered a quarter that shattered Wall Street forecasts, reporting revenue of $42.2 billion for the three months to 30 June – a 37% year-on-year leap. Analysts polled by LSEG had pencilled in growth of 31.2%, meaning the cloud division added nearly $2.5 billion more than the market expected. The immediate reaction was a 6% surge in Amazon’s stock in after-hours trading, on top of a 4% gain during the regular session.

The blockbuster performance was fuelled by an explosion in enterprise spending on artificial intelligence workloads. As companies race to train and deploy large language models, they are renting ever-larger slices of AWS compute, storage and custom silicon. The numbers provide the strongest evidence yet that the tech industry’s gigantic AI infrastructure build-out – which some analysts estimate will surpass $700 billion this year across the largest hyperscalers – is translating into tangible top-line growth.

The upbeat tone was tempered, however, by a sharp contraction in free cash flow. Amazon disclosed that trailing 12-month free cash flow stood at $7.6 billion, compared with $18.2 billion a year earlier, reflecting the relentless capital expenditure required to keep data centres humming and AI chips flowing. The results, which follow similarly strong cloud numbers from Microsoft and Alphabet, reinforce a narrative that AI demand remains extraordinarily robust, even as investors fret about a possible overbuild of capacity.

The AI Cloud Gold Rush: How AWS Is Capitalising on Demand

AWS Pulls Away from the Pack

A 37% growth rate at a unit already generating more than $42 billion a quarter is a remarkable feat that widens the gap with rivals. Microsoft Azure and Google Cloud both posted accelerating revenue, but neither came close to AWS’s scale or absolute dollar gains this quarter. By securing multi-year deals with AI heavyweights – OpenAI, Anthropic, Meta, Pinterest and Snowflake were all named this year – AWS is locking in the most critical workloads of the next decade while customers are still building their AI infrastructure. The commercial logic is clear: once a developer trains a model on a particular cloud’s custom chips and data pipelines, switching costs become enormous.

The Capex Calculus: Can Amazon Spend to Its Own Beat?

The free cash flow decline is the flipside of the AI opportunity. Amazon is spending billions on new data centres, specialised processors and energy supply, betting that the surge in AI usage will justify the outlay long before overcapacity becomes a real risk. The AWS beat goes a long way toward validating that bet, at least for now. If the growth rate slips back toward 30% or below in coming quarters, the stock’s current re-rating could quickly reverse. But as long as AI workloads are expanding at a triple-digit clip – earlier this year AWS disclosed its AI annualised revenue run rate had crossed $15 billion – the investment case holds up.

Prime Day and E-Commerce Provide a Safety Net

The retail side of the business continues to hum. This quarter’s Prime Day, held from 23 to 26 June, drove an estimated $26.4 billion in consumer spending according to Adobe Analytics, underscoring Amazon’s dominance in e-commerce even as it pours cash into its cloud future. The steady cash generation from retail and third-party seller services gives Andy Jassy’s company the luxury of funding its AI ambitions without having to strip the dividend or buybacks that shareholders often demand. That dual engine – a high-growth cloud unit paired with a mature, cash-producing commerce operation – remains Amazon’s strongest strategic asset.

What the AWS Beat Means for Investors and Cloud Customers

• Investors: Watch AWS’s Q3 growth rate closely; maintaining low-to-mid 30% growth would confirm that the $700bn industry capex wave is being absorbed. Another quarter above 35% could lift the entire big-tech complex.
• Cloud customers: The forthcoming wave of new data-centre capacity means more compute will become available later this year. Companies planning AI projects should lock in reserved-instance pricing now to hedge against potential cost increases as demand heats up further.
• Competitors: AWS’s deal pipeline – especially the inclusion of Meta, OpenAI and Anthropic – signals that even AI-first firms see value in outsourcing infrastructure rather than building it entirely in-house. That trend may pressure rivals’ own chip-development roadmaps.
• Suppliers and partners: Amazon’s appetite for custom silicon and networking hardware is set to expand. Chip designers and data-centre equipment makers with existing relationships with the company are likely to see orders accelerate in the second half of 2026.