AWS accelerates again, Amazon (AMZN.US) AI investment enters return verification phase

On July 31 Beijing time, after the U.S. stock market closed, Amazon released its Q2 2026 financial report. The most important signal from this report is not just the double-beat of revenue and profit, but that AWS has re-entered an acceleration phase, with AI and self-developed chip businesses beginning to provide quantifiable income to address the market's skepticism over massive capital expenditures.

According to Amazon’s official financial report, the company’s net sales for Q2 reached $200.6 billion, up 20% year-on-year; operating profit was $27.5 billion, up 43% year-on-year. Net profit reached $62.6 billion, with diluted earnings per share at $5.75, which includes $53.4 billion in non-operating pre-tax income from the Anthropic investment. Therefore, to judge the quality of Amazon’s current performance, more attention should be paid to four main lines: operating profit, AWS growth rate, advertising growth, and cash flow pressure.

I. Highlights: Growth no longer relies solely on retail, AWS becomes the core variable

From a business structure perspective, Amazon’s Q2 showed a combination of “steady retail, strong advertising, and accelerating cloud.” North American revenue was $116.2 billion, up 16% year-on-year, with operating profit of $9.1 billion; international revenue was $42.2 billion, up 15% year-on-year, with operating profit of $1.7 billion. The retail business did not drag down group profit; Prime Day being moved up to Q2, improvements in delivery efficiency, as well as growth in fresh and daily essentials, all contributed to the consumer-side performance.

The advertising business also maintained high prosperity. Amazon officially revealed that advertising revenue increased by 26% year-on-year. Driven by e-commerce traffic, Prime Video ads, and AI advertising tools, advertising has become an important force in enhancing the retail business’ gross margin structure. For investors, the significance of advertising lies in turning Amazon’s retail business from a low-margin transaction platform into a composite monetization model of “transaction + media + data.”

However, the real story-changer this quarter was AWS. AWS Q2 revenue was $42.232 billion, up 37% year-on-year, marking the fastest growth rate in 18 quarters; operating profit was $16.621 billion, up 64% year-on-year, with an operating margin of 39.4%. This means Amazon is not only ramping up in the AI cloud race, but also starting to show simultaneous income acceleration and margin improvement.

II. AI commercialization lands: AWS, Anthropic, and self-developed chips form a closed loop

In recent quarters, the market's main concern for Amazon has been: with Microsoft Azure and Google Cloud having a more prominent AI narrative, is AWS losing its leading edge? The Q2 report gives a direct answer. Amazon disclosed AWS currently has an annualized revenue run rate of $169 billion, backlog orders of $496 billion, and both AI business and chip business annualized revenue run rates exceeding $25 billion, all with triple-digit year-on-year growth.

This set of data has changed the framework for judging Amazon’s AI investments. Previously, the market saw only rapidly rising capital expenditures, but now, supporting evidence from the demand side is emerging. AWS’s AI business drives training, inference, data storage, vector databases, security governance, and CPU workloads; self-developed chips such as Trainium and Graviton help reduce customer costs and improve Amazon’s infrastructure efficiency. Amazon revealed in its report that leading AI labs like Anthropic and OpenAI have already committed to multi-year, multi-gigawatt levels of Trainium usage, granting higher visibility for AWS’s future revenue realization.

From an industrial logic perspective, Amazon’s strength lies not in stand-alone model capability but in the combination of cloud infrastructure, model platforms, enterprise data, chips, and application-layer services. In Q2, customer spending on Amazon Bedrock exceeded the sum of all previous quarters, indicating surging demand from enterprises for “multi-model, governable, deployable” AI platforms. Entering the inference era, applications like agents, code generation, and enterprise workflow automation will simultaneously consume GPU, CPU, storage, and database resources—precisely where AWS’s integrated infrastructure shines.

III. Capital expenditure is the point of contention: short-term cash flow pressure, long-term depends on order fulfillment

The biggest controversy for Amazon this quarter remains capital expenditures. The company raised its 2026 cash capital expenditure guidance from about $200 billion to around $220 billion. According to the Associated Press, CEO Andy Jassy said on the earnings call that even at $220 billion, the company cannot meet all of 2026 demand, and expects the supply-demand imbalance to persist into 2027, with notably robust reserved demand already for 2028.

High capex directly suppresses free cash flow. According to Amazon’s official report, for the 12 months through June 30, 2026, operating cash flow was $161.4 billion, up 33% year-on-year; but free cash flow was a cash outflow of $7.6 billion, compared with a cash inflow of $18.2 billion in the same period a year ago, mainly due to property and equipment acquisitions rising $66.1 billion, with related investment mainly aimed at AI infrastructure.

This is a common challenge for all major tech giants now: AI infrastructure requires spending upfront and confirming income later, with a gap caused by capacity build-up cycles and depreciation pressure. The difference for Amazon this quarter is that, with AWS accelerating growth, $496 billion in backlog orders, and annualized revenue run rates over $25 billion for AI and chip businesses, the company offers a relatively clear explanation for the input-output relationship. The market is willing to buy into this logic in the short term, but continued observation is needed to see if AI income growth can consistently outpace capex and depreciation pressure.

IV. Market response: From “burning cash concern” to “validating returns”

After the financial report was released, Amazon’s stock price strengthened significantly. MarketWatch reported that Amazon shares surged over 13% at one point post-earnings, potentially marking its best single-day performance in years; Investopedia cited market opinions pointing to AWS’s outperformance and signs of AI investment returns as key drivers for the stock price rally. It also noted JPMorgan raised Amazon’s price target to $365, citing enhanced AWS momentum and improved expectations for AI investment returns.

However, Q3 guidance still presents a short-term “speed bump.” Amazon expects Q3 net sales to be $197 to $202 billion, up 9% to 12% year-on-year; operating profit is expected to be $22.5 to $26.5 billion. The company explained that Prime Day moving from Q3 last year to Q2 this year affects the year-on-year comparison; excluding the Prime Day timing impact, Q3's revenue growth rate would be about 400 basis points higher; meanwhile, foreign exchange is expected to have a roughly 80-basis-point adverse effect.

Therefore, investment judgment on Amazon needs to be split into two layers: in the short term, the stock price has quickly reflected the Q2 beat, but Q3 revenue guidance and free cash flow pressure may introduce volatility. In the mid-to-long term, if AWS continues to accelerate and AI and chip revenue continue to expand, the market’s valuation anchor for Amazon may shift from being a “retail platform + cloud profit center” to an “AI infrastructure platform + global consumption ecosystem.”

V. Investment conclusion: Amazon returns to the AI cloud mainstream

In summary, the core highlight of Amazon’s Q2 earnings is that AWS is no longer merely a profit-maintaining cash cow but has once again become the group’s fastest-growing and most reliable engine. Retail business secures the base, advertising continues to enhance profitability, and AWS demonstrates with income growth, profit margin, and order backlog that AI demand is entering a realization phase.

Risks are also clear: capex is already at a historical peak, free cash flow has turned negative, and future depreciation pressure may gradually become more prominent. If AI client demand or enterprise inference applications grow below expectations, the market may re-examine the rationale for high capital investment. But at least for this quarter, Amazon delivered a report that helps ease “AI cash burn anxiety.” For long-term investors, key tracking metrics going forward should include AWS growth rate, annualized income from AI and chip businesses, changes in order backlog, the pace of capex, and free cash flow recovery progress.

VI. Earnings call summary: Key signals from management

• AWS growth reaccelerates. Management reported AWS grew 36.7% year-on-year in Q2, accelerating for the fifth consecutive quarter, and posting the fastest pace in 18 quarters, with sequential quarterly revenue up over $4.6 billion.

• AI and chip business become new growth drivers. Amazon disclosed annualized revenue run rates for both AWS AI business and chip business exceeded $25 billion, both achieving triple-digit year-on-year growth.

• Backlog orders provide income visibility. AWS backlog orders reached $496 billion, up triple digits year-on-year; management believes this reflects customers’ long-term commitment to computing power and cloud services.

• Capex continues upward. The company raised 2026 cash capital expenditure guidance to about $220 billion, mainly directed toward AWS and generative AI infrastructure, but also including robots, semiconductors, and satellites.

• Supply still insufficient. Andy Jassy said on the earnings call that even after the capex raise, the company still cannot meet all of 2026’s demand, and expects the supply-demand pattern to persist into 2027, with substantial demand already seen for 2028.

• Payback period explained in detail. Management emphasized that long-cycle assets like data centers, land, power, and buildings require advance investment, whereas short-cycle assets like servers and networking equipment are deployed as customer demand arises, meaning revenue recognition lags behind capital expenditure.

• Retail remains stable. Prime member delivery speed hit a record high, fresh and everyday essentials grew faster than overall business, and both North America and international businesses remained profitable.

• Advertising business continues rapid growth. Advertising revenue grew 26% year-on-year, with Prime Video ads, live sports, and AI ad tools cited as key growth drivers.

• Q3 guidance remains cautious. The company expects Q3 sales of $197-202 billion, operating profit of $22.5-26.5 billion; management noted Prime Day timing and foreign exchange will affect year-on-year performance.

• Core observation point. The central theme from the call is: AI investments are not short-term expenses but pre-built around clear orders and long-term computing needs; however, the market still needs to see the next few quarters to validate cash flow recovery and the return on capex.