Amazon delivered stronger-than-expected second-quarter earnings on Thursday, a performance that resonated positively with investors. Net sales climbed by 20%, with cloud revenue emerging as a particularly strong contributor. This combination of favorable results propelled Amazon's stock up by nearly 10% in after-hours trading.
Crucially, Amazon is demonstrating an unwavering commitment to data center spending, bucking the conventional investor sentiment that often advocates for companies to curtail such significant capital outlays.
A specific financial line item vividly illustrates Amazon's robust appetite for infrastructure investment. For the fiscal year concluding June 30, Amazon allocated $173 billion to property and equipment—a broad category encompassing GPUs, natural gas turbines, and land acquisitions—a substantial increase from the $107.65 billion spent in the preceding year.
Furthermore, the company revised its 2026 capital expenditure forecast upwards, from $200 billion to $220 billion. This increased projection comes despite Amazon beginning to draw on its cash reserves to cover costs. The quarter concluded with the company holding $7.6 billion less cash than it did 12 months prior, marking its first period of negative free cash flow this year.
Under typical circumstances, such rapidly expanding expenses would be a difficult proposition for investors to accept. However, Amazon possesses a powerful revenue engine that effectively justifies this spending. AWS revenue surged by 37% year-over-year, reaching $42 billion for the quarter. While this figure doesn't numerically balance the capital expenditure in a direct calculation, it emphatically indicates that demand is growing in parallel with supply. This trend offers considerable reassurance to investors, especially given the multi-year lead time between commencing data center construction and realizing its full capacity.
Significantly, Amazon's strategic involvement in artificial intelligence extends beyond the construction of massive data centers. The company is also making profound long-term investments in specialized chips, including the Trainium TPU and the Arm-based Graviton processor. While these projects do not typically appear in capital expenditure figures, they have the potential to meaningfully enhance the profit margins of the company's cloud business.
During the company's Q2 earnings call, CEO Jassy remarked, “We see the AI business following very much the same margin trajectory we saw in the core business before.” He further elaborated on Amazon’s approach, stating, “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.”
This dynamic is not unique to Amazon. Similar patterns were observed with Microsoft and Google, both of whose shares also climbed following reports of strong cloud revenue. Conversely, companies such as Meta, which are characterized by significant capital expenditure and a less defined revenue source, continue to face intense skepticism from investors. Meta's stock, for example, declined by 8% after its earnings release this week, as investors focused on its cash flow challenges and ongoing spending.
Naturally, investors prioritize revenue and are generally averse to expenses—this is a fundamental principle of market operations. Yet, it is vital not to overlook the broader implications for the AI economy. Currently, investors are positioning cloud hosts as the most dependable component of the AI stack, while maintaining skepticism regarding the underlying economic models for AI labs and burgeoning AI startups.
However, Amazon's hosting revenue directly translates into someone else's AI expenditure. In the specific case of Anthropic, this represents literally the same financial flow.
If the current spending levels by major AI labs and their clients prove to be unsustainable, the revenue stability for Amazon and other cloud hosts will inevitably be compromised. While genuine competition and differentiation exist at every layer of the technology stack, a failure of AI demand to sustain itself would undoubtedly lead to challenging times for all involved parties.
Ultimately, the entire discussion circles back to David Cahn’s $3 trillion question: Is there sufficient demand to justify this massive buildout, or is there not? Cloud-hosting services like AWS may operate a few steps removed from this core demand problem, but this distance does not render them immune to its potential impact.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.