Amazon gave Wall Street a clearer answer to the question hanging over the AI trade: cloud demand is accelerating, but the bill to build enough capacity is still rising.
The company said on July 30, 2026, that Amazon Web Services sales rose 37% from a year earlier to $42.2 billion in the second quarter, its fastest growth in 18 quarters. Overall net sales rose 20% to $200.6 billion, and operating income climbed to $27.5 billion.
The catch is cash flow. Amazon said free cash flow for the trailing 12 months swung to a $7.6 billion outflow, compared with an $18.2 billion inflow a year earlier, primarily because purchases of property and equipment increased for artificial intelligence infrastructure. That is the tradeoff investors are now being asked to price: faster AI-linked cloud demand today, with a larger upfront buildout before those data centers fully earn back their cost.
The numbers
AWS remained Amazon's profit engine. The cloud unit generated $16.6 billion in operating income, up from $10.2 billion in the same quarter last year, while North America operating income rose to $9.1 billion and international operating income reached $1.7 billion.
Net income jumped to $62.6 billion, or $5.75 a diluted share, but that headline profit included $53.4 billion of non-operating pre-tax other income, primarily tied to Amazon's investments in Anthropic. That makes operating income and cash flow more useful signals for judging the core business.
Amazon also raised the spending stakes. The Associated Press reported that the company increased its 2026 technology and AI capital-spending plan by $20 billion, to about $220 billion, after the stronger quarter. Investors appeared willing to accept the larger bill because AWS growth showed demand is not just theoretical.
Why investors and customers care
For investors, the report shifts the AI question from whether customers want more cloud capacity to how much cash Amazon must commit before the revenue arrives. A faster-growing AWS can justify bigger data-center, chip and power commitments, but negative free cash flow makes the timetable matter.
For business customers, the same numbers point to a capacity race. Amazon said its AI and chips businesses each exceeded $25 billion annual revenue run rates, and it described continued demand for Trainium chips, Bedrock models and agentic AI tools. That suggests more services are coming, but also that scarce infrastructure could shape pricing, availability and product priorities.
The caveat
This is not a simple earnings beat story. Amazon's results were helped by the Anthropic-related gain, and the company warned that foreign exchange, energy prices, tariffs, memory-chip costs and customer demand could affect future results. Heavy AI spending can look disciplined when cloud revenue accelerates, and excessive if demand cools.
What to watch next
The next test is whether AWS can keep growing fast enough to absorb the capital plan. Watch third-quarter sales guidance of $197 billion to $202 billion, AWS operating margin, property-and-equipment purchases, and any update on AI capacity constraints. The market's message for now is narrow: Amazon got credit for growth, but the AI spending meter is still running.
