Amazon reports Q2 AWS revenue up 37% YoY to $42.2B, above estimates of a 31.21% increase, and AWS operating income up 64% to $16.6B; AMZN jumps 13%+
and still won't have enough capacity to meet demandIsabella Simonetti /Wall Street Journal:Amazon Shares Jump as Cloud Sales—and Spending—AccelerateJordan Novet /CNBC:Amazon's AWS posts fastest growth since 2021, citing AI and chip demandSpencer Soper /Bloomberg:Amazon Gains After Fifth Quarter of Cloud Sales Growth
Context & Ripple Effects
AWS’s growth had been running at a materially slower pace in the preceding coverage: Q2 2025 revenue growth was 17.5%, while its Q4 2024 operating-income growth outpaced sales growth. The latest result marks a sharp acceleration in both demand and profitability.
The report also pairs stronger cloud sales with faster spending and an explicit capacity shortfall. That makes the result relevant not just as an earnings beat, but as evidence that available AI and chip capacity remains a binding constraint for AWS.
First-order effects
- AWS gains immediate financial momentum: revenue exceeded expectations and operating income grew faster than sales, while Amazon’s after-hours share move reflects investors’ reassessment of the cloud unit’s near-term earnings power.
- Customers seeking AWS capacity face continued allocation constraints, since AWS says current infrastructure will not meet demand; Amazon must keep expanding supply while demand is elevated.
Second-order effects
- The combination of accelerating sales and unmet demand reinforces pressure on rival cloud platforms to secure and deploy AI and chip capacity, rather than relying on more measured buildouts.
- Higher infrastructure spending flows through to the AI-compute supply chain, but constrained capacity can also limit how quickly cloud customers can shift workloads or launch compute-intensive services.
Third-order effects
- If sustained, this points to a cloud market in which access to power, data centers and specialized compute increasingly determines growth, making capital deployment a more important competitive differentiator than incremental feature releases.
- The key uncertainty is duration: sustained demand would validate a longer infrastructure investment cycle, while demand that normalizes before new capacity arrives could compress returns on that spending.
The trend: This is one data point in the AI infrastructure supercycle, where cloud demand is accelerating faster than providers can add specialized compute capacity.