Amazon surpassed expectations in its second-quarter earnings report on Thursday, resulting in a surge in investor confidence. With a 20% increase in net sales, the company’s cloud revenue proved particularly impressive, leading to a nearly 10% rise in its stock during after-hours trading.
Notably, Amazon continues to increase its investment in data centers, contradicting the expectations of some investors who might prefer a more conservative approach.
A significant indicator of Amazon’s commitment to expanding its infrastructure can be seen in its substantial spending. The company allocated $173 billion for capital expenditures in the fiscal year ending June 30, a sharp rise from the previous year’s $107.65 billion. This expenditure covers critical areas such as GPUs, natural gas turbines, and land acquisitions.
Moreover, Amazon has revised its capital expenditure forecast for 2026 upwards from $200 billion to $220 billion, even as it starts to use its cash reserves to fund these costs. As a result, at the end of the quarter, Amazon had $7.6 billion less cash compared to the same time last year, marking its first quarter of negative free cash flow this year.
Typically, rising expenses can be troubling for investors. However, Amazon’s steady revenue growth supports these expenditures. The company’s AWS segment saw a remarkable 37% year-over-year increase, generating $42 billion for the quarter. While this revenue might not fully offset the capital expenses, it indicates that supply is keeping pace with growing demand. Given the lengthy time it takes to develop a data center and bring its capacity online, this growth offers reassurance to investors.
Amazon’s strategy is not solely focused on large data centers. The company is making significant long-term investments in technologies such as the Trainium TPU and the Arm-based Graviton processor. These initiatives may not directly factor into capital expenditures but are expected to enhance profit margins for its cloud services considerably.
“We see the AI business following a similar margin path to what we observed in our core operations previously,” said CEO Andy Jassy during the earnings call. “AWS and Amazon Bedrock can thrive without a single predominant model, as no single model will dominate the market.”
This trend is mirrored by similar outcomes at Microsoft and Google, which also experienced stock increases following robust cloud revenue reports. Conversely, companies like Meta, which have high capital expenditures without a clear revenue stream, continue to face skepticism from investors. Meta’s stock dropped by 8% after its quarterly earnings announcement, highlighting concerns about its cash flow and ongoing expenditures.
Investors generally favor revenue growth and dislike rising costs—it’s a fundamental aspect of market behavior. However, it is crucial to recognize the larger implications within the AI sector. Currently, investors regard cloud service providers as the most dependable segment of the AI ecosystem, while maintaining doubts about the financial viability of AI labs and startups.
Moreover, Amazon’s cloud revenue can be viewed as a direct cost for AI companies like Anthropic.
If the spending by major labs and their clients proves unsustainable, it would disrupt revenue stability for Amazon and its cloud peers. There is intense competition and differentiation across the entire AI stack; if the demand for AI solutions wanes, it could lead to negative consequences for all involved.
Ultimately, the situation boils down to a critical question regarding demand: Is there sufficient demand to support this infrastructure expansion? While AWS may not be immediately impacted by this demand issue, it is not completely shielded from its effects.
