AMZN surged 15%—is this a turning point or an early overdraw?

NEXT PICK · Market Insights

AMZN surged 15%—is this a turning point or an early overdraw?

Friday, July 31, 2026

A company with a market capitalization of $2.9 trillion rose 15.32% in a single trading day, with its market value increasing by about $390 billion in a single day, marking its best single-day performance in 14 years—this isn't the script of a micro-cap monster, but rather the answer sheet Amazon delivered on July 31, 2026.

Closing price $271.58, trading volume about 2.6 times the three-month daily average, and the stock price is about 15% above the 200-day moving average. On the surface, the price-to-earnings ratio is only 21.85 times, which is ridiculously cheap for a giant whose revenue is accelerating.

But if you flip the financial report to the cash flow statement, you'll see another number: over the past twelve months, free cash flow has been a net outflow of $7.6 billion, compared to a year ago at $18.2 billion in net inflow on the same scale. With the same financial report and two different narratives, why is the market applauding?

Latest price $271.58 ▲ +15.32%

AWS Quarterly Revenue $42.2 billion ▲ +37%

TTM free cash flow -$7.6 billion ▼ From positive to negative

It's not retail that is rising, but the saying, "Clouds can go even faster."

Breaking down this 15.32% increase, the real trigger is just one number: AWS's quarterly revenue was $42.2 billion, up 37% year-over-year—the fastest in 18 quarters—while the market had previously expected it to be around 31%.

The renewed acceleration of cloud business means that the question that previously weighed on the entire sector—"When will AI investment actually turn into revenue?"—has been temporarily answered. AWS posted an operating profit of $16.6 billion for the quarter, with an operating margin of 39.4%. While accelerating, margins continue to expand.

The group level is equally impressive

Quarterly revenue was $200.6 billion, surpassing $200 billion for the first time in a single quarter in history; Operating profit was $27.5 billion, up 43% year-on-year; Operating margin rose from 11.4% in the same period last year to 13.7%. Eight sellers collectively raised their target prices after the earnings reports.

But there is a detail here that must be clarified

Quarterly net profit was $62.6 billion, with earnings per share of $5.75, both impressive figures, including $53.4 billion in non-operating pre-tax income, mainly from the company's investment appreciation in Anthropic.

In other words, that seemingly cheap 21.85x P/E ratio is stuffed into the denominator a book unrealized gain that doesn't appear every quarter. What truly represents the main company's self-sustaining capability is the $27.5 billion operating profit, not the $62.6 billion net profit.

The company walks on three legs, with the third leg growing into the main torso

Amazon's revenue structure is now clearly divided into three parts: North American retail reached $116.2 billion in the quarter, up 16% year-over-year; international retail was $42.2 billion, up 15%; AWS was $42.2 billion, up 37%.

Interestingly, the distribution of profits is completely asymmetric to revenue—North American retail contributed $9.1 billion in operating profit, international retail $1.7 billion, while AWS, which is just as large as international retail, contributed $16.6 billion in just one quarter.

There is also a hidden thread that is often overlooked

Advertising revenue for the quarter was $19.8 billion, up 26% year-over-year. This money is embedded in retail statements, yet it has profit margins close to software, which is the key to sustaining the retail segment's profit margins.

So, what exactly is this company? Rather than calling it an e-commerce platform, it's more accurate to say it's a three-tiered structure: the bottom layer is self-developed chips and data centers, the middle layer is cloud services and logistics networks, and the top layer is the shopping app you know.

Implaceability grows from the bottom. Once enterprise customers bring databases, AI training pipelines, and compliance audits into AWS, the cost of migration goes beyond just rewriting APIs, including retraining operations teams and redoing compliance certifications—this is an engineering cost, but an organizational cost.

Self-developed chips have raised this wall even higher. Inference loads optimized for Trainium cannot be transferred directly to NVIDIA GPUs or Google TPUs; Graviton is already used by 98% of the top 1,000 EC2 customers, with revenue commitments growing nearly threefold quarter-over-quarter.

A P/E ratio of 21.85 times—is it cheap or just a distraction?

First, break down the number that is most likely to mislead. TTM's earnings per share of $12.43 included the unrealized profit from the $53.4 billion Anthropic investment that quarter; After excluding such one-off gains, the actual operating P/E ratio is significantly higher than the surface value of 21.85 times.

The rest of the valuation anchors vary in opinion

A price-to-sales ratio of less than 4 times is not expensive for a giant with growth close to 20%; But the enterprise value multiples are close to 29 times, an absolute level that already prices AWS's high growth for the coming years.

As for free cash flow yield, which is usually the strongest anchor, it simply can't be used right now—free cash flow over the past twelve months is negative $7.6 billion, and the numerator is negative, so this ratio itself loses its meaning.

Cash flow is the sharpest contradiction in this financial report. TTM's operating cash flow was $161.4 billion, up 33% year-on-year, with no issues in self-sustaining capacity; The problem lies on the spending side: TTM capital expenditure was $169 billion, a year-on-year surge of 64%. The company stated in its filing that this "mainly reflects investment in artificial intelligence."

Management also raised its full-year 2026 capital expenditure guidance from $200 billion to about $220 billion, citing rising memory prices as one reason. On one hand, income is accelerating; on the other, investment is accelerating. So, who will last longer in this race?

Has the good news materialized? My judgment is: the AWS acceleration line has already been fully priced and slightly overshot within a 15% single-day increase; As for the capital expenditure return cycle, the market has simply chosen not to ask further.

From the 'Year of Efficiency' to the 'Year of Armaments,' the market has just confirmed a shift in gears

At this stage, Amazon has already passed the commercialization validation phase and entered a breakthrough phase of scale. There are three layers of judgment: revenue growth accelerated from 12.4% for the full year FY2025 to 20% this quarter, with profit growth outpacing revenue growth and operating leverage being released.

Self-developed chips have also crossed a critical point—the annualized revenue run rate of both AI and chip businesses has surpassed $25 billion, and the chip business has maintained triple-digit year-on-year growth; Graviton5 is now fully commercialized, with computing performance 25% higher than Graviton4, and nearly twice the growth rate of Graviton4 that year.

Technically, the market is also on the bulls' side

The stock price is $271.58, above the 50-day moving average of $246.38 and also above the 200-day moving average of $234.75, which is in a bullish alignment. However, the deviation rate from the 200-day moving average of about 15%, along with relative strength indicators near the overbought zone, all indicate that some short-term momentum has been depleted.

The second half of the arms race is not about share, but about endurance

On the cloud and AI infrastructure track, Amazon faces three competitors of equal scale. Microsoft's 2026 capital expenditure guidance is about $175 billion. Azure is deeply integrated with OpenAI, leading the application layer of generative AI (the Copilot ecosystem), and enterprise customers have a naturally smooth migration path from Office 365 to Azure.

Alphabet's capital expenditure guidance is $195 billion to $205 billion, the highest among the four, leveraging TPU-dedicated chips for hardware differentiation in training and inference, and Gemini can also collaborate with search advertising; The shortcoming is that cloud business share still lags behind AWS, enterprise penetration is low, and it faces antitrust pressure.

Meta's $130 to $145 billion capital expenditure mainly serves its own model training and ad recommendations, not external cloud services. Its conflict with Amazon is more about AI talent and upstream procurement of memory and chips, rather than customer competition.

NVIDIA's position is most nuanced: it is both Amazon's most important upstream supplier and extending into cloud services through DGX Cloud. Amazon's strategy is the vertical integration of Trainium and Graviton—both reducing procurement costs and locking customers into its own chip ecosystem.

The shape of the moat can be recorded like this

Amazon isn't defending itself with a single moat; instead, it has dug five at once—switching costs, bilateral network effects, economies of scale, vertical integration, and compliance certification barriers.

The flywheel on the retail side is still turning

Millions of third-party sellers bring massive products, and these massive products attract hundreds of millions of Prime members. Member traffic boosts the return on ad spending, and ad revenue feeds back into logistics and technology investment.

It's just that the nature of the competition has changed. When the combined capital expenditure of the four giants approaches or even exceeds $600 billion in capital expenditure by 2026, the key to victory will no longer be who grabs two more points of market share, but whose cash flow can last longer.

The ceiling is high enough, but the road to the ceiling requires burning money first

Based on addressable market estimates, the global public cloud market is expected to reach approximately $800 billion to $1 trillion by 2026. As the leading player, AWS will serve about 30% to 35% of global cloud infrastructure spending. AWS's current annualized revenue run rate is about $169 billion, indicating still room for penetration.

The stronger evidence is the order book

AWS backlog has reached $496 billion. This means that revenue visibility in the coming years will not depend on the imagination of new demand, but on whether capacity can be delivered on schedule—management has made it clear that even if capital expenditure is raised to $220 billion, it will still not meet all 2026 needs.

Retail and advertising space has also not peaked

Global e-commerce transaction volume is expected to exceed $7 trillion by 2026, with Amazon's penetration rate around 10%; The advertising business generated nearly $80 billion in annualized revenue, with its share in the global digital advertising market still in single digits to just over 10%, indicating rapid expansion.

Each of the five bull cards has financial backing

The logic of the bulls is not based on imagination, but on the numbers already realized this quarter:

▲ Bull Case

The six question marks for bears point to cash flow

Bears do not deny that this quarter was beautiful; they question the sustainability and cost of this beautiful moment:

▼ Bear Case

In the next three quarters, four factors will determine whether this repricing process can hold its ground

This round of rally has already fulfilled optimistic expectations ahead of schedule. The next key point is not "whether things will get better," but "whether it can hold on":

•	Q3 2026 Earnings Report: Validated revenue guidance of $197 to $202 billion and AI capital expenditure return rhythm, with a relatively high impact
•	After Graviton5 is fully commercialized, AWS gross margin improvement and customer migration progress can be observed around Q4 2026, with a neutral to positive impact
•	Whether self-developed chips can maintain triple-digit growth and break through the $25 billion annualized run rate ceiling will be verified throughout FY2026, with a relatively high impact
•	Whether the release of CPO/NPO supply chain capacity can ease AI data center network bottlenecks is a timeline pointing to 2027, which is neutral

It only rose 16% in a year, but dropped more than half in a single day

Looking at the long term, AMZN's return over the past 12 months is about 16%, and over the past 30 days, it's been 11.1%—meaning the year's gains have been almost entirely swallowed up by the most recent month, especially this day.

In fact, there have been two price resets over the past year. The first time will occur between late 2025 and early 2026: FY2025 full-year revenue was $716.9 billion, up 12.4% year-over-year. The market confirmed the company had returned to growth from the "efficiency year" of 2022-2023, but the stock price reaction was mild, described as "steady but not exciting."

The second time was from July 30 to 31. AWS growth jumped to 37%, operating profit increased by 43%, capital expenditure was raised to $220 billion, and management said "still not enough"—these three pieces of information combined directly dispelled previous valuation concerns about AI investment.

It's also interesting to check afterwards which expectations were met and which were revised: revenue acceleration and profit margin expansion materialized, and self-developed chips reached a $25 billion annualized run rate from concept to $25 billion; Negative free cash flow is something most sell-side models have not fully accounted for before.

⚠️ Risk Notice • Free cash flow has turned negative, with $220 billion in annual capital expenditure supported by bond issuance • Rising memory chip prices and supply fluctuations directly drive up construction costs, restricting the pace of capacity delivery • Net profit includes a one-time investment gain of $53.4 billion, with apparent valuation indicators severely distorted • Q3 revenue guidance growth fell to 9%-12%, below market expectations • After a single-day surge of 15.3%, the technical outlook is close to overbought and short-term pullback pressure is clear

🟡 Neutral Cloud acceleration is real, but the cash flow cost is just as real as the overbought position.

💬 Discussion Bottom line: Fundamentals have improved, prices are high, better to wait for pullbacks than to chase highs. For specific buy and sell positions, post AMZN on the official WeChat account to check.

Data source

•	Source: NextPick real-time market snapshot + Amazon Q2 2026 earnings report and official announcements + SEC filing documents + mainstream financial media public coverage.