Behind Amazon (AMZN.US)'s two-day jump of 20%: AWS accelerates growth and verifies AI demand capital expenditure and depreciation pressure raises concerns
The Zhitong Finance App learned that Amazon (AMZN.US) stock prices closed higher for two consecutive trading days after announcing second-quarter results that exceeded expectations, with a cumulative increase of about 20%. According to the bulls, despite the 2026 capital expenditure forecast being raised and free cash flow turned negative, the accelerated growth of Amazon's cloud computing business revenue for the fifth consecutive quarter mitigated concerns that its huge investment in artificial intelligence (AI) capital would not return. However, there are still cautious market participants that believe that for Amazon, what has actually been verified is artificial intelligence (AI) demand, not the relationship between construction costs and actual economic returns, which means that Amazon's future stock price may be once again pressured by capital expenditure.
Amazon's Q2 performance exceeds expectations, and the cloud business accelerates growth to verify AI demand
According to financial reports, Amazon's total revenue for the second quarter increased 20% year over year to 2006 billion US dollars, better than analysts' average forecast of 1970 billion US dollars; operating profit was 27.5 billion US dollars, up 43% year over year; net profit was 62.6 billion US dollars, up 245% year over year, including 53.4 billion US dollars of non-operating income before tax, mainly from its investment in Anthropic; diluted earnings per share were 5.75 US dollars, far higher than analysts' average expectations of 1.82 billion US dollars.
The core Amazon Cloud Services (AWS) revenue increased 37% year over year to $42.2 billion, better than analysts' average forecast of $40.6 billion, and the fastest growth rate since the fourth quarter of 2021. Investors paid close attention to the growth of AWS before the financial report was released. Amazon CEO Andy Jassi said that AWS is “booming,” and pointed out that its AI and self-developed chip business divisions are growing strongly. Both have annualized revenue of more than 25 billion US dollars, and both have achieved three-digit percentage growth over the same period last year.
Like other big tech companies, Amazon is investing massively in data centers and chips to seize the opportunities brought by the rapid growth in demand for AI and cloud computing services. Amazon's capital expenditure in the second quarter reached $54.2 billion, up from $32.1 billion in the same period last year. Its heavy investment in AI products and infrastructure led to a negative free cash flow — as of the end of the second quarter, the company had a net free cash outflow of $7.6 billion over the past 12 months, compared to a net inflow of $18.2 billion a year ago.
Amazon also raised its 2026 capital expenditure forecast from the previous forecast of $200 billion to $220 billion. Jaxi said that most of this spending will go to the AI field. Jasi said that the rise in memory prices has boosted expectations of its capital expenditure. He added that Amazon's spending boom is unlikely to slow down in the short term. “Even at this level, we will not be able to have enough production capacity to meet all demand by 2026, and I believe this trend will continue in 2027. In fact, we think the demand for 2028 is already very significant”.
Jasi also said, “We are in a very good position in this wave of AI transformation.” He emphasized that Amazon's investment in AI infrastructure is critical to meet the growing demand for its cloud services. He said that AWS's backlog of work this quarter (that is, contract projects that have not yet gone online) has reached 496 billion US dollars.
The market is happy to see AI requirements verified but ignores capital expenditure and depreciation concerns
Amazon's second-quarter earnings report further raised the level of AI demand verification — in fact, this verification was already reflected in Amazon's first-quarter earnings report — and strengthened confirmation of related trends. This is also an important reason for the sharp rise in Amazon's stock price after its performance.
At the same time, however, the rise in Amazon's stock price seems to have selectively overlooked the fact that the company raised its annual capital expenditure expectations. If we refer to Amazon's own past development history, capital expenditure pressure is generally not conducive to the company's stock price performance, and the current capital investment cycle is very different from the past. The scale of this round of capital expenditure is higher, and as the company continues to announce new capital investment plans, the return on investment implementation phase is constantly being delayed. Meanwhile, the gap between the growth rate of capital expenditure and the rate of revenue growth is still very clear.
Amazon raised its capital expenditure plan from about 200 billion US dollars to about 220 billion US dollars in the second quarter earnings report. The estimated time from investment to first generating bill revenue is 6 to 24 months, and some assets have a useful life span of up to 30 years. This means that the return on investment will gradually be realized within a few years (not several quarters), and as demand stabilizes, different assets will generate returns at different rhythms in different economic environments. Additionally, when Amazon's previous capital expenditure cycle of 2021-2022 began, the company had a stronger free cash flow base, which is completely different from the current situation.
This is why some market participants think Amazon's stock price will fall again after the initial market excitement brought about by the second-quarter earnings report subsides. According to some investors, the strong demand for AI shown by Amazon's AWS business performance does not answer the question of economic return on capital expenditure.
Some analysts point out that the strong performance of the AWS business does not prove that Amazon's current capital investment is reasonable. AWS is just one link in the AI investment chain—it shows that companies are migrating their systems to the cloud, that's all. It's actually on the weakest side of AI investment.
The really important capital expenditure issues are the AI training business and custom chips (Trainium), and the demand and economic models for these two fields are still in their early stages. Although the data currently published by the AI training business and chip business is performing well, demand is highly concentrated in a few AI laboratories. More importantly, the profit model of these businesses is still unverified. The market still doesn't know. What kind of profit margins can these businesses achieve in a stable state of operation? What is the return on capital under ongoing depreciation pressure? What is the actual useful life of the assets in question?
Some signs are even more worrying. In addition to the fact that free cash flow has turned negative, depreciation costs have also continued to rise. In the second quarter, Amazon's overall depreciation and amortization expenses (D&A) increased year-on-year from about US$15.2 billion to approximately US$20 billion. Among them, depreciation expenses related to AWS rose from approximately US$4.8 billion to approximately US$8.1 billion. More importantly, depreciation is growing faster than revenue growth — AWS depreciation and amortization expenses increased by around 67% in the second quarter, while revenue growth was only about 37%.
This shows that Amazon is in a highly aggressive AI infrastructure construction phase. These inevitable costs in the construction process will have a negative impact on the company's cash flow and earnings report (and the stock price should reflect this), although these factors themselves do not necessarily mean that the company has major problems.
Furthermore, the approximately $90 billion in assets added in the first half of the year has only just begun to be depreciated, so profit margins for the next few quarters will face greater depreciation burdens. The company also faces a conflict between the five-year lifetime assumption of servers and the risk of rapid decommissioning of AI devices.
Overall, the increase in Amazon's stock price after the announcement of the second-quarter earnings report may have exceeded the valuation increase brought about by AI demand verification, while negative factors, including increased capital expenditure and depreciation pressure, do not seem to have been fully priced by the market. The market is ignoring the pressures and uncertainties that exist in Amazon's capital investment phase. Investors may continue to wait and see until the AI demand and economics indicated by Amazon's capital expenditure are proven, and the gap between AWS revenue growth and depreciation cost growth begins to narrow before re-evaluating buying opportunities.