Amazon reported better-than-expected second-quarter results on Thursday, prompting its stock to rise nearly 10% in after-hours trading. The company’s net sales grew 20%, and Amazon Web Services (AWS) stood out with revenue up 37% year over year to $42 billion for the quarter.
Large infrastructure spending continues
Amazon is not dialing back spending on data centers. For the fiscal year ended June 30, the company spent $173 billion on property and equipment — a category that includes GPUs, natural gas turbines, and land — up from $107.65 billion the previous year. It also raised its 2026 capital expenditures guidance from $200 billion to $220 billion.
Alongside higher planned capex, Amazon’s cash position has weakened: the company ended the quarter with $7.6 billion less cash than it had 12 months earlier, and this quarter marked its first period of negative free cash flow this year.
Revenue helps justify the outlays
Rapidly rising expenses would normally concern investors, but Amazon has a strong revenue engine in AWS. The $42 billion in quarterly AWS revenue helps explain the heavy investment. While that income does not fully offset capex in raw arithmetic, it demonstrates that demand is increasing alongside supply — a reassuring sign given the multi-year lag between breaking ground on a data center and selling its capacity.
Chip projects and margin implications
Amazon’s AI strategy goes beyond building large data centers. The company is also investing in custom chips, including the Trainium TPU and the Arm-based Graviton processor. These projects aren’t always reflected directly in capex figures but can materially improve cloud-margin economics.
"We see the AI business following very much the same margin trajectory we saw in the core business before," Jassy said on the company’s Q2 earnings call. "AWS and Amazon Bedrock can have a wildly successful business without its own frontier model, and the reason is that there’s not going to be a single model to rule them all."
Industry context and risks
The pattern of strong cloud revenue and investor enthusiasm has appeared at other large cloud hosts too: Microsoft and Google also saw shares rise after reporting robust cloud revenue. By contrast, companies with heavy capex but less clear near-term revenue — such as Meta — faced investor skepticism; Meta’s stock fell about 8% after its quarterly report this week amid concerns over cash flow and continued spending.
It’s also important to note that cloud-hosting revenue ultimately reflects other companies’ AI bills. In Anthropic’s case, for example, the hosting payments are literally the same money flowing to cloud providers. If spending by big AI labs and their customers proves unsustainable, hosting revenue will not be stable for Amazon and its peers. There is competition and differentiation at every layer of the stack, but if AI demand softens, the fallout could be broad.
The central question
Everything ties back to what David Cahn framed as the "$3 trillion question": is there sufficient demand to justify this infrastructure buildout, or not? Cloud-hosting services like AWS may sit a few steps removed from end-user demand, but they are not immune to its fluctuations.
