Amazon.com (AMZN.US) shares surged nearly 20% over two trading sessions following a better-than-expected second-quarter earnings report, marking a sharp rally driven by strong cloud computing growth.

For bullish investors, the standout highlight was AWS revenue accelerating for the fifth consecutive quarter, easing fears that massive artificial intelligence investments would fail to generate returns. However, some cautious analysts argue that what has truly been validated is AI demand itself, not the economic link between construction costs and actual returns, suggesting the stock could face renewed pressure from rising capital spending.

Where to begin

Amazon reported total second-quarter revenue of $200.6 billion, up 20% year-over-year and surpassing the average analyst estimate of $197 billion. Operating income reached $27.5 billion, a 43% increase. Net income soared to $62.6 billion, up 245%, which included a $53.4 billion pre-tax non-operating gain from its investment in Anthropic. Diluted earnings per share came in at $5.75, far exceeding the consensus estimate of $1.82.

The most critical metric—AWS revenue—grew 37% year-over-year to $42.2 billion, beating expectations of $40.6 billion and marking its fastest growth rate since the fourth quarter of 2021. CEO Andy Jassy described AWS as "thriving," noting that its AI and custom chip business segments each generate an annualized revenue run rate exceeding $25 billion, both growing at triple-digit percentages year-over-year.

Like other major tech companies, Amazon is aggressively investing in data centers and chips to capitalize on surging demand for AI and cloud services. Capital expenditures in the second quarter hit $54.2 billion, up from $32.1 billion in the same period last year. This heavy spending on AI products and infrastructure pushed free cash flow into negative territory: the trailing twelve-month free cash flow was a net outflow of $7.6 billion, compared to an inflow of $18.2 billion a year earlier. Amazon also raised its 2026 capital expenditure outlook to $220 billion, up from a previous estimate of $200 billion, with Jassy noting that the majority of this spending will be directed toward AI.

Jassy attributed the higher spending expectations partly to rising memory prices and said the spending spree is unlikely to slow in the near term. "Even at this level, we won't have enough capacity to meet all demand by 2026, and I believe this trend will continue into 2027. In fact, we already see significant demand for 2028," he added. Jassy emphasized that AWS's backlog of unlaunched contracts has reached $496 billion, underscoring the critical role of AI infrastructure investments in meeting growing cloud service needs.

Why just 10 ASX 200 shares?

The second-quarter earnings further validated AI demand—a trend already visible in the first quarter—and provided stronger confirmation of its trajectory, fueling the post-earnings stock surge. However, the rally seems to have selectively ignored the company's upward revision to its full-year capital spending forecast. Historically, capital spending pressures have not been favorable for Amazon's stock performance, and this current cycle differs significantly from past ones.

The scale of this spending cycle is much larger, and the timeline for investment returns continues to be pushed back as new spending plans are announced. The gap between capital expenditure growth and revenue growth remains stark. The company raised its capital spending plan from approximately $200 billion to around $220 billion. The time from investment to initial billing revenue is estimated at 6 to 24 months, with some assets having useful lives of up to 30 years. This means returns will materialize gradually over years—not quarters—and different assets will generate revenue at different paces under varying economic conditions.

Additionally, when Amazon initiated its previous capital spending cycle in 2021-2022, it had a much stronger free cash flow base, a situation that has completely reversed. Some market participants believe that once the initial post-earnings excitement fades, the stock could retreat. In their view, the strong AWS performance does not resolve the question of whether the current capital spending is economically justified. AWS is just one link in the AI investment chain—it shows that companies are migrating systems to the cloud, but it sits at the weaker end of AI investment impact. The truly critical capital spending issues lie in AI training and custom chips (Trainium), where demand and economic models remain in early stages. While these segments show promising data, demand is heavily concentrated among a few AI labs, and their profit models remain unproven. It is unclear what profit margins these businesses can achieve under steady-state conditions, what return on capital looks like under continuous depreciation pressure, or the actual useful lives of the assets.

Some indicators are even more worrying. Besides free cash flow turning negative, depreciation expenses continue to rise. In the second quarter, Amazon's overall depreciation and amortization expense grew from approximately $15.2 billion to $20.0 billion year-over-year. Specifically, AWS-related D&A rose from about $4.8 billion to $8.1 billion. More importantly, depreciation growth is outpacing revenue growth—AWS D&A grew roughly 67% in the second quarter, while revenue grew only 37%. This indicates Amazon is in a highly aggressive phase of building AI-related assets. These costs naturally impact cash flow and reported profits, factors the stock should reflect, though they do not necessarily indicate fundamental problems. Furthermore, the roughly $90 billion in new assets added in the first half of the year have only just begun to be depreciated, meaning subsequent quarters will face even greater depreciation burdens. The company also faces a conflict between its five-year server useful life assumption and the rapid obsolescence risk of AI equipment.

In summary, the post-earnings stock surge may have already priced in the valuation uplift from AI demand validation, while negative factors—including increased capital spending and depreciation pressure—appear not yet fully priced in by the market. The market is overlooking the pressures and uncertainties inherent in Amazon's current investment phase. Investors may choose to wait on the sidelines until Amazon's capital spending-driven AI demand and its economics are proven, and until the gap between AWS revenue growth and depreciation expense growth begins to narrow, before reassessing buying opportunities.