Amazon Earnings: AWS Growth Hits 37% as Apple Outlook Disappoints

Last updated: August 1, 2026, 10:52 a.m. CEST (4:52 a.m. ET). Market prices refer to the July 31, 2026 U.S. closing session unless otherwise stated.

Amazon delivered the clearer winner from the latest round of megacap technology earnings. The company reported second-quarter net sales of 0.6 billion, up 20% from a year earlier, while Amazon Web Services grew 37% to .2 billion and consolidated operating income rose 43% to .5 billion. The speed of the AWS acceleration, the cloud unit’s 39.4% operating margin and management’s assertion that future capacity is already heavily reserved gave investors something they had been demanding from the artificial-intelligence spending cycle: visible revenue and profit arriving alongside the investment bill.

Apple’s fiscal third quarter was strong by ordinary standards and impressive for a company of its size. Revenue rose 16% to $109.4 billion, diluted earnings per share increased 29% to $2.02, iPhone sales advanced nearly 22% and Mac sales climbed almost 29%. Yet Apple shares fell sharply because the market was looking beyond the completed quarter. Services revenue came in below consensus expectations, management’s September-quarter revenue-growth forecast trailed Wall Street’s target, and supply constraints raised questions about how much demand Apple can fulfill without sacrificing margin or raising prices.

The contrast is more instructive than a simple “Amazon beat, Apple missed” summary. Apple did not miss the quarter as a whole. Amazon’s reported net income also requires careful handling because a large non-operating valuation gain connected principally to Anthropic inflated earnings per share. The durable message lies deeper in the statements: AWS is growing faster while generating substantial operating profit; Amazon’s retail and advertising operations are also expanding; cash investment is consuming free cash flow; and Apple’s product cycle is powerful but increasingly exposed to component availability, memory costs and expectations that leave little room for an imperfect outlook.

Key Takeaways

• Amazon’s central result: Net sales reached $200.6 billion, up 20%, and operating income reached $27.5 billion, up 43%, in the quarter ended June 30, 2026.
• AWS was the decisive number: AWS revenue rose 37% to $42.2 billion, its fastest growth in 18 quarters, while AWS operating income increased 64% to $16.6 billion.
• The earnings-per-share caveat: Amazon’s $62.6 billion net income and $5.75 diluted EPS were boosted by $53.4 billion of non-operating income, driven mainly by valuation adjustments related to Anthropic investments.
• The spending trade-off: Amazon’s trailing-12-month operating cash flow rose to $161.4 billion, but free cash flow fell to an outflow of $7.6 billion because property-and-equipment spending rose sharply, primarily for AI infrastructure.
• Apple’s completed quarter was strong: Revenue rose to $109.4 billion and diluted EPS reached $2.02, with iPhone, Mac and Services setting June-quarter revenue records.
• Apple’s outlook caused the concern: Management forecast 9% to 11% revenue growth for the September quarter, below the approximately 12% Wall Street target reported by Reuters, while describing meaningful supply constraints.
• The market verdict was decisive: Amazon closed July 31 at $271.58, up 15.25%, while Apple closed at $308.91, down 7.14%.
• What matters next: Amazon must turn unprecedented capital spending into sustained AWS growth without allowing negative free cash flow to become structural. Apple must convert strong demand into shipments while controlling component costs and protecting its premium margins.

What Happened in Amazon’s Second Quarter

Amazon’s quarter combined three developments that rarely arrive together at the same intensity: rapid top-line growth, expanding operating profit and a sharp acceleration in the company’s highest-value business. According to Amazon’s earnings release, net sales increased from $167.7 billion in the second quarter of 2025 to $200.6 billion in the second quarter of 2026. Foreign exchange contributed only about $0.1 billion, so the 20% reported growth rate was not meaningfully dependent on currency translation.

The operating result was stronger than the revenue result. Operating income rose from $19.2 billion to $27.5 billion. That lifted the consolidated operating margin from roughly 11.4% to 13.7%. A two-percentage-point-plus improvement in operating margin is substantial for a company processing more than $200 billion of quarterly sales across retail, logistics, advertising, subscriptions and cloud infrastructure. It suggests that the quarter was not merely a Prime Day volume event or an accounting benefit from exchange rates. Amazon converted a greater share of revenue into operating profit.

The three operating segments all remained profitable. North America sales rose 16% to $116.2 billion, and segment operating income increased 21% to $9.1 billion. International sales rose 15% to $42.2 billion, while operating income increased to $1.7 billion from $1.5 billion. AWS sales also totaled approximately $42.2 billion, but the economics were very different: AWS generated $16.6 billion of operating income, compared with $1.7 billion for the International segment on nearly the same revenue.

That comparison shows why investors focus so heavily on cloud growth. AWS represented about 21% of Amazon’s quarterly revenue but supplied roughly 61% of consolidated operating income. North America remained a meaningful profit contributor, and International continued its multi-year transition from loss-making expansion to positive earnings. Still, AWS carried the group’s profit mix.

Amazon’s business-line data also showed broad participation. Online stores revenue increased 15% to $70.4 billion. Third-party seller services grew 16% to $46.8 billion. Advertising services advanced 26% to $19.8 billion, and subscription services rose 12% to $13.7 billion. Those categories matter because Amazon is no longer best understood as an online retailer with a cloud side business. Its economics increasingly depend on services layered around commerce and computing: marketplace commissions, fulfillment fees, sponsored listings, subscriptions, cloud capacity, databases, AI tools and chips.

The quarter therefore answered several questions at once. Demand for AWS accelerated. The cloud unit retained very high profitability even as Amazon expanded infrastructure. Retail growth did not disappear under the weight of a larger base. Advertising outgrew the company. International operations remained profitable. The concern shifted from whether Amazon could generate enough demand to justify its investment program to whether it can finance and execute that program without creating a lasting cash-flow deficit.

Amazon Earnings Scorecard: Reported Results and Their Meaning

Source: Amazon’s second-quarter 2026 earnings release and Form 10-Q. Dollar amounts are reported figures. Percentages are company-reported or calculated from reported figures.

Why 37% AWS Growth Changed the Market Conversation

The most important number was not Amazon’s total revenue, net income or after-hours price move. It was AWS growth. The cloud segment expanded 37% year over year, up from 28% growth in the preceding quarter, and reached an annualized revenue run rate of about 9 billion. Amazon described it as the fastest AWS growth in 18 quarters.

Cloud growth rates have become a proxy for the commercial adoption of generative AI because the largest model developers, software companies and enterprises require immense amounts of computing, networking and storage. The connection is not perfect. AWS revenue includes traditional workloads, databases, cybersecurity, analytics, migration projects and ordinary storage. Yet the acceleration indicates that demand is broad enough to absorb a rapidly expanding infrastructure base.

The result also changed the interpretation of Amazon’s position relative to Microsoft Azure and Google Cloud. During earlier phases of the generative-AI cycle, Microsoft appeared to have the clearest narrative advantage because of its partnership with OpenAI and the rapid integration of AI features into Azure and Microsoft software. Google had its own models, custom tensor-processing units and deep research organization. Amazon’s story looked more fragmented: Bedrock as a model-access layer, Anthropic as a strategic partner, Trainium and Inferentia as custom chips, and an enormous infrastructure build whose revenue payoff was not always immediately visible.

AWS’s second-quarter acceleration did not settle the competitive question, but it weakened the claim that Amazon was structurally behind. A 37% growth rate on a quarterly base above billion is not a small-company rebound. It implies more than $11 billion of additional quarterly revenue compared with a year earlier. The company also produced a 39.4% AWS operating margin, up from 32.9% in the year-earlier quarter. That combination—faster growth and higher margin—is what made the report unusually persuasive.

There are several plausible drivers. Customers are expanding AI training and inference workloads. Enterprises are modernizing older systems. Capacity constraints may be easing enough for AWS to recognize revenue that demand had already supported. Custom silicon can lower the cost of certain workloads and improve the economics of capacity. Long-term contracts can improve visibility, though they can also include pricing concessions. Amazon’s Form 10-Q states that AWS sales growth primarily reflected increased customer usage, partly offset by pricing changes driven mainly by long-term customer contracts.

The qualification matters. Revenue acceleration is not solely a price story. Usage rose enough to overcome contractual pricing effects. That is a stronger signal of demand than growth produced mainly by higher unit prices. It also suggests that the economic value of new workloads is reaching the income statement rather than remaining only in order books, memoranda of understanding or promised future deployments.

AWS Is Amazon’s Profit Engine, Not Merely Its Growth Engine

Revenue growth attracts attention, but the profit contribution explains the valuation. AWS generated .6 billion of operating income in the quarter, up from .2 billion a year earlier. Its operating margin reached 39.4%. On approximately one-fifth of Amazon’s consolidated revenue, AWS produced about three-fifths of operating income.

The arithmetic changes how the rest of Amazon should be read. Retail creates enormous scale, traffic, fulfillment density, merchant relationships and consumer data. Advertising monetizes that commercial traffic. Prime subscriptions deepen customer loyalty and support purchase frequency. AWS supplies a disproportionate amount of operating profit. The businesses reinforce one another, but their financial roles differ.

AWS’s margin was not achieved by avoiding investment. Amazon’s technology and infrastructure expense increased 22% in the quarter to $33.2 billion. The filing attributes the increase primarily to higher infrastructure spending, including depreciation and amortization. AWS operating income still grew 64%. That means the current revenue and efficiency benefits were large enough to absorb substantial expense growth.

Investors should not assume that a near-40% segment margin will rise indefinitely. Cloud margins can move with equipment life assumptions, depreciation schedules, energy costs, chip availability, customer mix, price reductions, utilization and the pace at which new data centers are filled. New capacity often carries costs before it reaches mature utilization. AI infrastructure can also have different depreciation and power profiles from older cloud workloads. A quarter of strong margin does not eliminate those sensitivities.

Even so, the Q2 result provides evidence that Amazon’s infrastructure is not simply a capital sink. AWS is currently converting demand into profit at a scale few businesses can match. The issue is no longer whether AWS can make money. It is how much capital must be committed to preserve growth, how durable the margin is, and whether custom chips and software services can keep Amazon’s returns attractive as competition intensifies.

The strongest interpretation is that AWS has entered a new demand cycle in which AI workloads add to, rather than replace, conventional cloud migration and consumption. The skeptical interpretation is that exceptional growth reflects a period of scarce capacity and urgent customer spending that may normalize after the largest model builders complete their initial buildouts. Both views can coexist. The quarter favors the first, but a single period cannot establish the long-term slope.

The $5.75 EPS Headline Needs an Accounting Footnote

Amazon reported net income of $62.6 billion, or $5.75 per diluted share, compared with $18.2 billion, or $1.68 per share, a year earlier. Those figures are accurate under U.S. generally accepted accounting principles. They are not a clean representation of the company’s recurring quarterly operations.

The reason appears in the income statement and investment disclosures. Amazon recorded $53.4 billion of “other income, net” in the quarter. The 10-Q says the gain was primarily related to upward valuation adjustments for nonvoting preferred stock in Anthropic and reclassification gains connected with convertible-note investments that converted into preferred stock. The company recorded approximately .5 billion of upward adjustments to Anthropic preferred stock in the quarter.

That does not make the gain fictitious. Amazon’s investments can create real economic value, and a funding round or other observable transaction can support a higher carrying value. It does mean that the resulting EPS is sensitive to valuation events that are outside the ordinary sale of goods, advertising or cloud services. The gain is also non-cash in the period. It can reverse if valuation conditions change, and its tax consequences differ from operating profit.

This is why operating income is the more useful headline for evaluating the quarter. Operating income excludes the large investment mark and rose 43% to .5 billion. That result remains exceptional without relying on the Anthropic valuation. AWS operating income, North America profitability and the International segment’s positive contribution all support the operational case.

It would also be misleading to subtract the entire $53.4 billion from net income and present the remainder as a definitive “adjusted profit.” Taxes, investment classifications and other non-operating items complicate that shortcut. The disciplined approach is to separate the measures rather than manufacture a new one. GAAP net income shows the total accounting result. Operating income shows the profit produced before interest, taxes and major non-operating valuation effects. Cash flow shows how much cash the business generated and reinvested.

The distinction has practical importance for valuation. A price-to-earnings multiple based on a quarter containing a large mark-to-market gain can look artificially low. Analysts and investors often normalize such items when estimating forward earnings. That does not guarantee a particular valuation conclusion; it simply prevents an unusual investment remeasurement from being mistaken for recurring cloud or retail profit.

Amazon’s AI Investment Bill: $220 Billion and Rising

The quarter’s strongest result arrived with the company’s largest financial risk. Chief Executive Andy Jassy said Amazon expects 2026 capital expenditures of about 0 billion, up from an earlier plan of 0 billion. The increase reflects higher infrastructure requirements and component costs, including memory, as well as investment in data centers, chips, robotics, satellites and the fulfillment network. The majority of technology-infrastructure spending supports AWS growth.

The scale is difficult to overstate. The Associated Press reported that Amazon spent roughly 8 billion in capital expenditures during 2025. A $220 billion plan would be about 72% higher. It would also exceed the annual economic output of many countries and represent more than the company’s entire Q2 revenue. Capital expenditure is not recognized as an immediate expense in the income statement; the assets are capitalized and depreciated over time. The cash leaves sooner than the accounting expense appears.

That timing explains the free-cash-flow reversal. Amazon’s trailing-12-month operating cash flow rose 33% to $161.4 billion. Purchases of property and equipment, net of proceeds and incentives, increased to $169.0 billion. Free cash flow therefore fell to an outflow of $7.6 billion, compared with an inflow of $18.2 billion a year earlier.

Quarterly cash capital expenditures were $53.1 billion, up from $31.4 billion in Q2 2025. For the first six months of 2026, they reached $96.3 billion, compared with $55.6 billion in the corresponding 2025 period. Those figures show that the elevated annual plan is not distant guidance; spending is already occurring.

Negative free cash flow is not automatically a sign of financial distress. Amazon generated more than $160 billion of operating cash flow over the preceding 12 months, retained access to debt markets and is investing in assets intended to produce future revenue. A company can rationally accept negative free cash flow when incremental returns are expected to exceed its cost of capital and when demand visibility is high.

The risk lies in the word “expected.” Data centers take time to build. Power connections, memory chips, networking equipment and advanced accelerators can be scarce. Customer demand can shift. AI training economics can improve, reducing the computing required for a given task. Competitors can cut prices. Large customers can design their own chips or diversify among providers. Assets may be technically current when ordered and less attractive by the time they are deployed.

Amazon’s answer is that much of the capacity is being built against identifiable demand. Reuters reported that management said a majority of available AWS capacity for 2027 and some capacity for 2028 had already been reserved. That is a meaningful reassurance, though reservation does not eliminate execution risk. Contract duration, cancellation provisions, pricing, customer concentration and the capital required to serve commitments all affect the eventual return.

Why the Market Accepted Amazon’s Spending and Rejected Other AI Bills

Big Tech’s 2026 capital-spending cycle has created a new market rule: investors are willing to tolerate extraordinary expenditure when the company can show a direct path from investment to revenue and margin. They are less forgiving when monetization is distant, indirect or difficult to measure.

Amazon’s quarter met that test. AWS revenue grew 37%. AWS operating income rose 64%. Segment margin reached 39.4%. Management described constrained supply and reserved future capacity. The spending increase therefore appeared connected to customer demand that was already producing revenue.

This does not mean the market has become indifferent to cash flow. Amazon’s negative trailing free cash flow is a real deterioration. It means investors judged the deterioration to be purposeful rather than accidental. The distinction can change quickly. If AWS growth slows while capital expenditure remains near 0 billion, the same spending program that supported the July rally could become the central criticism.

The comparison with other megacap companies is useful but should not be simplified. Meta’s spending supports recommendation systems, advertising efficiency, model development and long-term platform ambitions. Alphabet invests in Google Cloud, search infrastructure, Gemini models and other products. Microsoft invests in Azure and its OpenAI relationship. Their returns arrive through different income statements and timelines. A dollar of capital expenditure is not directly comparable across companies.

Still, Amazon gave investors the clearest current-period pairing of infrastructure growth and segment profit. Reuters described a growing divide in market tolerance for AI spending and reported that Amazon’s result reassured investors after concerns about the sector’s rising bill. The July 31 price action reflected that distinction: Amazon gained 15.25%, while several other large technology stocks had recently faced harsher reactions to higher spending plans.

The market’s response also illustrates how expectations dominate absolute performance. Amazon’s $220 billion plan is objectively huge. If the company had reported 20% AWS growth and weaker margins, the plan might have been interpreted as evidence of declining capital efficiency. At 37% growth and 39.4% margin, it looked more like a race to add capacity in front of demand. The facts are the same; the operating context changes their meaning.

Custom Chips, Bedrock and the Economics of AI Capacity

Amazon’s AI strategy is built around more than rented Nvidia accelerators. AWS offers access to third-party models, its Bedrock service, general-purpose cloud infrastructure and custom chips including Trainium for training and Inferentia for inference. The economic objective is to control more of the stack and reduce dependence on the most expensive external components.

Custom silicon can improve economics in several ways. It may lower the cost per unit of computation for workloads optimized to use it. It can reduce exposure to shortages. It gives AWS another way to differentiate pricing and performance. It can deepen customer dependence on Amazon’s software and development tools. It also requires large design budgets, software support and consistent customer adoption.

Jassy said Amazon’s AI and chip businesses had each surpassed annualized run rates of billion, according to the company’s release and subsequent reporting. Run-rate figures are not the same as audited annual revenue categories. They extrapolate a current level and can include businesses defined by management. Even so, the statement provides scale context: Amazon is no longer describing AI as a small experimental line buried inside AWS.

Anthropic is central to the strategy. Amazon has invested heavily in the model developer and has commercial arrangements involving AWS cloud services and AWS chips. The 10-Q describes additional preferred-stock investments, contractual obligations related to AWS chip performance and a financing facility tied to delivery milestones and a future Anthropic liquidity event. The relationship combines customer, supplier, investor and strategic-partner roles.

That alignment has advantages. Anthropic can become a large anchor customer for AWS infrastructure and custom chips. Amazon gains exposure to a leading model developer. AWS customers gain model access through Bedrock. The arrangement can help Amazon demonstrate that Trainium is capable of serving demanding frontier-model workloads.

It also introduces complexity. Investment valuation gains can distort net income. Capital commitments and commercial contracts can create financial exposure beyond ordinary cloud sales. Concentration in a few large model developers may make demand appear broader than it is. If model economics or competitive positions change, the value of both the investment and associated infrastructure demand can change.

The most relevant measure over the next several quarters will be not a single AI announcement but the relationship among AWS growth, backlog conversion, segment margin, depreciation and capital expenditure. Custom chips should eventually show up in better unit economics, stronger capacity availability or greater customer adoption. If they do not, the strategic narrative will outrun the financial evidence.

Retail Was Not a Passenger in the Quarter

AWS dominated the discussion, but Amazon’s retail operations were also strong. North America sales rose 16% to $116.2 billion. Segment operating income increased to $9.1 billion, and the operating margin held near 7.9%. International sales increased 15% to $42.2 billion, with a 4.1% operating margin.

Those numbers reflect years of work on fulfillment-region design, inventory placement, shipping speed and cost control. Amazon has reorganized parts of its U.S. network to place products closer to customers, reducing the distance and number of handling steps required for delivery. Faster service can improve conversion and purchase frequency while also reducing certain transportation costs when inventory is positioned correctly.

The trade-off is that speed requires inventory, local capacity and operational precision. Shipping costs rose from .4 billion to .9 billion in the quarter. Fulfillment expense increased 14% to $29.6 billion. Amazon’s filing says the higher fulfillment cost reflected sales growth and investment in the network, partly offset by operational efficiencies.

Prime Day timing also mattered. Amazon held the event in the second quarter of 2026 rather than the third quarter, shifting revenue between periods. That makes the 20% Q2 growth rate stronger and the 9% to 12% Q3 guidance weaker than a simple year-over-year comparison would imply. Amazon explicitly stated that Q3 growth would be nearly four percentage points higher if the Prime Day effect were excluded from both years.

The timing adjustment does not erase the retail strength. North America grew 16%, online stores rose 15%, third-party seller services increased 16% and subscriptions grew 12%. Prime Day helped, but the business was not dependent on one category. Advertising and seller services expanded alongside merchandise sales.

International profitability deserves separate attention. Amazon spent years building markets that often produced losses. A 4.1% operating margin remains below North America’s, but consistent positive profit changes the portfolio. It reduces the extent to which domestic retail and AWS must subsidize international expansion and creates room for local investment without automatically depressing consolidated earnings.

Retail’s strategic value also extends beyond its own margin. The marketplace generates product searches, purchase data and advertiser demand. Fulfillment services make the platform more useful to merchants. Prime increases engagement. Those activities create the audience and transaction context that Amazon’s advertising business monetizes.

Advertising Is Becoming a Third Profit Pillar

Amazon advertising services revenue rose 26% to $19.8 billion. That made advertising nearly half the size of AWS revenue in the quarter and larger than many stand-alone media businesses. Amazon does not disclose advertising operating profit separately, but the business is generally understood to have attractive incremental economics because it monetizes traffic already present on the company’s properties.

The key advantage is commercial intent. A consumer searching for a product on Amazon is often closer to purchase than a user casually browsing a social feed or watching entertainment. Sellers and brands can pay to appear in sponsored results, product pages, display placements and video inventory. Amazon can connect exposure with transactions occurring in its marketplace, subject to privacy, attribution and regulatory constraints.

Advertising growth supports North America and International segment profitability because ad revenue is reported within those segments rather than as a separate operating segment. The filing says increased advertising sales helped drive higher operating income in both regions. This partly explains how retail margins can improve even while shipping and fulfillment expenses rise.

The opportunity is broader than sponsored listings. Amazon owns streaming properties, devices and live sports rights, and it can sell video advertising across those surfaces. It can also offer tools to advertisers using its demand-side platform. The combination of commerce data and media inventory makes the business an increasingly important competitor to Google, Meta and traditional television advertising.

There are limits. More advertising can damage the shopping experience if organic results become difficult to find. Regulators may scrutinize how marketplace data is used. Brands may question incrementality if ads merely redirect customers who would have bought anyway. Competitive ad platforms can improve attribution or offer lower prices.

For now, 26% growth suggests Amazon has not reached an obvious saturation point. The segment is important because it diversifies the profit story. If AWS growth moderates in a future quarter, advertising can still support operating margin. If retail volume slows, higher-value services can cushion the effect. Amazon’s earnings quality improves when multiple service businesses contribute rather than leaving the company dependent on merchandise volume alone.

Amazon’s Q3 Guidance: Weaker Than the Headline Quarter, Better Than It First Appears

Amazon forecast third-quarter net sales of $197 billion to $202 billion, representing growth of 9% to 12%. The midpoint, $199.5 billion, was below the roughly $203.9 billion analyst expectation cited by the Associated Press. Operating income is expected to be $22.5 billion to $26.5 billion, compared with $17.4 billion in the third quarter of 2025.

The revenue range created the main point of caution in the immediate earnings reaction. Growth of 9% to 12% is much slower than Q2’s 20%. The comparison is distorted by Prime Day, which occurred in Q2 this year and Q3 last year. Amazon said the growth rate would be nearly 400 basis points higher after excluding Prime Day from both periods. On that basis, the underlying range would look closer to approximately 13% to 16%, before considering the 80-basis-point expected foreign-exchange headwind.

The operating-income guide is more constructive. Its midpoint is $24.5 billion, about 41% above the prior-year quarter’s $17.4 billion. The low end still implies substantial growth. That suggests management expects continued profit expansion even as reported revenue comparisons normalize.

Guidance remains conditional. Amazon lists energy prices, trade policies, memory-chip availability, customer spending, interest rates, labor markets and geopolitical conditions among the factors that can affect results. The company also assumes no additional acquisitions, restructurings or legal settlements.

Investors will need to separate three questions when Q3 arrives. First, did underlying retail growth remain healthy after adjusting for event timing? Second, did AWS maintain an elevated growth rate as the comparison base increased? Third, did operating profit remain strong while depreciation and infrastructure expenses rose?

A quarter near the low end of sales guidance would not automatically invalidate the Q2 story if AWS and operating income remain strong. Conversely, a sales result near the high end could still disappoint if it comes with weaker cloud margins or a larger cash-flow drain. The quality and mix of growth matter more than a single consolidated figure.

How the Stock Market Interpreted Amazon’s Report

Amazon shares closed at $271.58 on July 31, 2026, up $35.92, or 15.25%, after trading as high as $273.20. The move added hundreds of billions of dollars to the company’s market capitalization and ranked among Amazon’s strongest one-day performances in years.

The size of the rally reflects more than an earnings beat. Investors revised several assumptions simultaneously: AWS may be reaccelerating rather than merely stabilizing; Amazon’s AI infrastructure is generating current revenue; custom chips and strategic model partnerships may be strengthening the platform; operating margins can expand despite high investment; and retail plus advertising are providing additional support.

Price action is evidence of changing expectations, not proof that those expectations will be met. A 15% gain raises the future standard. Once a stock reprices for sustained 30%-plus cloud growth and successful AI monetization, ordinary execution may no longer be enough. The next disappointment can produce an equally sharp reversal.

The market also looked through the negative free-cash-flow figure. That choice is rational only as long as investors believe capital expenditure is building productive assets against durable demand. The Q2 result supplied that belief. Future quarters must renew it.

At the index level, Amazon’s gain helped the Nasdaq and S&P 500 finish higher despite Apple’s decline. Reuters reported that the S&P 500 gained 0.70%, the Nasdaq rose 1% and the Dow added 0.53% on July 31. The divergent megacap moves showed that “AI exposure” was no longer enough as a broad category. The market distinguished between companies on the basis of revenue conversion, guidance and spending efficiency.

The result also affected suppliers and adjacent infrastructure names. Strong hyperscaler demand can support companies selling accelerators, networking gear, power equipment, cooling systems, memory and servers. That does not mean every supplier benefits equally. Amazon’s custom-chip strategy can reduce dependence on some third-party components, while capital constraints or architecture changes can shift spending among vendors.

Apple’s Quarter Was Stronger Than Its Stock Reaction Suggested

Apple reported fiscal third-quarter revenue of $109.4 billion for the period ended June 27, 2026, up 16% from $94.0 billion a year earlier. Net income rose to $29.8 billion from $23.4 billion, and diluted EPS increased to $2.02 from $1.57. The company described the result as its strongest June quarter for revenue and EPS.

The product mix was powerful. iPhone revenue rose 21.7% to $54.3 billion. Mac revenue increased 28.7% to $10.4 billion. Wearables, Home and Accessories revenue grew 6.5% to $7.9 billion. iPad revenue fell 5.9% to $6.2 billion. Services revenue increased 12.1% to $30.7 billion.

Geographic performance was broad. Americas revenue rose to $45.8 billion from $41.2 billion. Europe increased to $29.4 billion from $24.0 billion. Greater China rose to $18.8 billion from $15.4 billion. Japan and the rest of Asia Pacific also grew. Apple said its installed base of active devices reached a record across major product categories and geographic segments.

Gross margin reached 50.1%, and operating income increased to $35.7 billion from $28.2 billion. Research and development expense rose 32% to $11.7 billion, indicating a substantial increase in product and technology investment. Apple’s profitability remained extraordinary: operating income equaled roughly one-third of revenue.

The quarter was not weak. The negative market reaction came from three issues: some of the upside was helped by tariff refunds, Services revenue missed expectations, and guidance suggested supply constraints would prevent Apple from fully converting demand into sales in the September quarter.

Why Apple Shares Fell Despite Record Revenue and EPS

Apple closed July 31 at $308.91, down $23.80, or 7.14%. The decline was not a judgment that the June quarter had failed. It was a repricing of the next quarter and the quality of the beat.

First, Apple said tariff refunds added approximately two percentage points to the 50.1% gross margin and $0.11 to diluted EPS. Excluding that EPS benefit, the quarter would have produced about $1.91 per share. Reuters reported that this still exceeded the $1.89 Wall Street estimate, so the underlying result remained sound. The refund nevertheless made the headline comparison look stronger than recurring operations alone.

Second, Services revenue of $30.74 billion missed the approximately $31.22 billion consensus cited by Reuters. Services carries strategic importance beyond its share of sales. It tends to have higher gross margins than hardware, provides recurring revenue and monetizes Apple’s installed base. A miss in that category can outweigh a hardware beat because investors assign considerable value to the stability and profitability of the ecosystem.

Third, Chief Financial Officer Kevan Parekh forecast September-quarter revenue growth of 9% to 11%, below the approximately 12% target reported by Reuters. He projected iPhone growth in the mid-teens, compared with a 17.6% market expectation, and gross margin of 47% to 48%.

Fourth, management described significant supply constraints involving advanced chipmaking capacity and memory. Chief Executive Tim Cook framed the problem as insufficient flexibility to meet strong demand, not a shortage of customers. That is preferable to weak demand, but the financial consequence can still be missed revenue, higher component costs and pressure to change pricing.

Finally, Apple entered the report with a demanding valuation and a strong share-price run. When expectations are elevated, a company can beat the completed quarter and still fall if the outlook does not preserve the same trajectory. The market was not comparing Apple with an average business. It was comparing the guidance with a valuation that assumed exceptional execution.

Apple’s Services Miss: Small in Dollars, Large in Meaning

Services revenue increased by roughly $3.3 billion year over year to $30.7 billion. A 12.1% growth rate on that base is substantial. The disappointment arose because analysts expected more and because Services has become central to the argument that Apple can grow beyond device replacement cycles.

The category includes the App Store, cloud services, payments, subscriptions, advertising, licensing and media offerings. The mix is not disclosed in enough detail to determine precisely which subcategories caused the shortfall. That limits interpretation. A lower-than-expected quarter could reflect timing, regional mix, regulatory changes, subscription trends or other factors.

Services matters for three reasons. It is recurring. It deepens switching costs by embedding users in storage, payments, applications and content. It can also produce a higher gross margin than hardware, helping the company preserve profitability when component costs rise.

The category is not risk-free. App Store economics face regulatory and legal pressure in multiple jurisdictions. Search-distribution payments can be affected by antitrust remedies. Consumers can reduce discretionary subscriptions. Competitive services can weaken Apple’s take rate or force product changes. The installed base provides a large opportunity, but regulators increasingly challenge the rules governing access to that base.

The Q3 result did not establish a structural slowdown. Services still set a June-quarter record. It did show that the market expects the category to do more than grow at a low-double-digit rate. When hardware delivers unusually strong growth, investors may also ask whether some services demand was pulled forward or whether monetization is keeping pace with device adoption.

Apple’s next reports should be judged on both absolute Services growth and the drivers beneath it. A return above expectations would reduce concern. Repeated misses would force a reassessment of the recurring-revenue premium embedded in the stock.

Supply Constraints and the “Strong Demand” Problem

Management’s explanation for Apple’s softer outlook was supply, not demand. Cook told Reuters that advanced chipmaking capacity was a major constraint in the June quarter and said Apple was evaluating options for memory suppliers. The company expects significant constraints to continue.

A demand-driven shortage is better than excess inventory, but it can still damage financial performance. Unfilled orders defer or lose revenue. Expedited components and alternative sourcing can raise cost. Product availability can shift sales among regions or models. Customers may wait, switch products or buy from competitors. If Apple raises prices to protect margin, unit demand can weaken.

Memory is particularly important because it affects many products and has experienced a sharp industry price cycle. Apple’s scale gives it purchasing power, but even a large buyer cannot instantly create advanced capacity. Long supply agreements can improve availability while locking in prices that later look high. Diversifying suppliers can reduce concentration but requires qualification and quality control.

The 47% to 48% gross-margin forecast indicates an expected decline from the tariff-refund-assisted 50.1% quarter. Some normalization was inevitable because the refund contributed about two percentage points. The remaining question is how much pressure comes from memory costs, product mix and supply inefficiency.

Apple can respond through pricing, product mix, supplier negotiations, design changes and operating expense discipline. Each option has trade-offs. Higher prices protect margin but test elasticity. Favoring premium models can increase average selling prices but limit volume. Redesigns take time. Supplier changes can introduce execution risk.

The market’s concern is therefore not simply that Apple lacks components. It is that a constrained company with a premium valuation may be unable to translate unusually strong demand into the revenue growth investors expected. The problem is operational rather than existential, but the valuation makes operational imperfections expensive.

Amazon and Apple Represent Two Different AI Business Models

The earnings split also exposed two distinct ways to monetize artificial intelligence. Amazon sells infrastructure. Apple sells devices, software and an ecosystem in which AI can improve the user experience. The revenue timing and capital structure are different.

Amazon’s AI revenue can appear directly through AWS consumption, reserved capacity, managed services and custom chips. When customers train or run models, usage can translate into cloud revenue. The company must spend heavily before and during that growth to build data centers and secure power, networking, memory and processors.

Apple’s AI monetization is less direct. Better on-device intelligence can stimulate upgrades, strengthen retention, support premium pricing and increase engagement with services. The company may not report a separate “AI revenue” line. Benefits can appear in iPhone sales, Mac sales, Services revenue or gross margin over time.

The second-quarter comparison favored Amazon because the market could see a direct link: 37% AWS growth, 64% AWS operating-income growth and a large capacity build. Apple’s quarter showed powerful hardware demand, but Services missed expectations and supply constraints limited the outlook. Its AI progress therefore had less immediate financial visibility.

That does not establish that infrastructure is the superior long-term model. Cloud providers face enormous depreciation, power and technology risks. Device ecosystems can monetize innovation through high-margin products without matching hyperscaler capital expenditure. Apple generated $35.7 billion of operating income on $109.4 billion of revenue in the quarter, a level of profitability that Amazon’s consolidated business does not approach as a percentage of sales.

The better conclusion is that the market currently rewards measurable AI revenue more than strategic optionality. Amazon provided measurable infrastructure growth. Apple provided strong product demand and a longer-term ecosystem argument, but its immediate guide introduced friction. Future product launches or services acceleration could reverse the relative narrative.

The Capital-Efficiency Debate Is Now the Central Big Tech Debate

For much of the past two years, the market asked whether AI demand was real. That question is becoming less useful. Revenue at major cloud providers, model-developer funding, data-center construction and semiconductor demand all indicate substantial activity. The more difficult question is whether the spending produces acceptable returns.

Capital efficiency depends on several variables: revenue per dollar of infrastructure, gross margin, utilization, asset life, energy cost, financing cost, customer concentration and the speed at which new hardware becomes obsolete. A company can report fast revenue growth and still destroy value if the assets required to generate it are too expensive or short-lived.

Amazon’s Q2 evidence was favorable but incomplete. AWS margin expanded. Operating income rose faster than sales. Future capacity was reportedly reserved. Those points suggest attractive utilization and demand. Free cash flow turned negative, showing that the capital burden is already larger than operating cash generation on Amazon’s chosen definition.

The debate will intensify as depreciation catches up with spending. Cash capital expenditures affect free cash flow immediately. Depreciation affects operating profit over the useful life of the assets. A rapid build can therefore produce a period in which operating margins look strong while cash flow weakens, followed by later periods in which depreciation rises. Revenue must continue expanding to absorb that expense.

Changes in estimated useful lives can also affect reported profit. Extending server life lowers annual depreciation; shortening it raises depreciation. Investors should read accounting-policy disclosures and not treat operating margin as independent of asset-life assumptions.

Amazon’s strongest defense is demand visibility. If reserved capacity converts to revenue and custom chips lower unit costs, the spending can generate durable returns. The strongest concern is technological and financial rigidity: 0 billion of annual investment creates a large installed base that must remain productive through fast architecture changes.

What the Results Mean for Nvidia, Super Micro and the AI Supply Chain

The video’s market discussion connected Amazon and Apple with infrastructure names including Nvidia and Super Micro Computer. The relationship is economically relevant but not one-directional.

Amazon’s capex increase supports the broad AI supply chain. Data centers require accelerators, general-purpose processors, memory, networking, storage, racks, power management, backup systems and cooling. Construction also benefits electrical-equipment and engineering providers. Strong AWS demand suggests that hyperscaler orders are tied to active workloads rather than speculative capacity alone.

Nvidia benefits from industry demand for advanced accelerators, but Amazon is simultaneously developing Trainium and Inferentia. Custom chips can expand total capacity while reducing the share captured by Nvidia in selected workloads. The outcome depends on performance, software compatibility, customer preference and supply. A growing AWS can purchase more third-party accelerators and deploy more custom chips at the same time.

Super Micro and other server vendors can benefit from rapid deployment cycles, but hyperscalers often design systems internally and source components through multiple channels. Revenue can be volatile, margins can be lower than those of chip designers, and customer concentration can be high. A hyperscaler capex announcement should not be translated mechanically into a specific supplier’s earnings.

Apple’s supply constraints send a different signal. Strong demand for advanced manufacturing and memory can support component pricing, but shortages can reduce device shipments. Suppliers may gain pricing power while customers face margin pressure. The same industry bottleneck can therefore be positive for one part of the chain and negative for another.

For the broader market, the two reports confirm that component availability has become a strategic variable. It affects cloud capacity, consumer-device production, capital spending and gross margins. Companies with secure supply and differentiated designs may gain share. Companies dependent on a narrow set of suppliers face greater volatility.

Valuation After the Earnings Split

Amazon’s 15.25% rally and Apple’s 7.14% decline changed the starting point for any valuation discussion. A strong report can improve fundamentals while also making the stock less forgiving. A weak reaction can reduce expectations while leaving the underlying company profitable and growing.

Amazon’s reported price-to-earnings ratio requires caution because trailing net income includes large investment valuation gains. Operating-income growth, normalized earnings and cash-flow expectations are more informative. The company’s valuation now assumes that AWS can sustain elevated growth, that capital expenditure will produce acceptable returns and that retail plus advertising will continue supporting profit.

Apple’s valuation remains tied to ecosystem durability, pricing power, installed-base monetization and capital returns. The quarter showed strong demand, but the stock’s premium requires confidence that supply constraints are temporary and Services can meet higher expectations. A company can be operationally excellent and still deliver weak shareholder returns if purchased at a price that assumes perfection.

Neither conclusion can be reduced to a simple earnings multiple. Amazon’s accounting earnings are affected by investment marks, and its free cash flow is depressed by a deliberate capital cycle. Apple’s margins were helped by tariff refunds, and its next-quarter revenue is constrained by supply. Scenario analysis is more useful than a single ratio.

For Amazon, the favorable scenario combines AWS growth above 30%, stable high-30s segment margins, continued advertising growth and improving free cash flow after the build. The adverse scenario combines slowing cloud growth, persistent $200-billion-plus capex, rising depreciation and lower utilization.

For Apple, the favorable scenario assumes component constraints ease, strong iPhone and Mac demand persists, Services reaccelerates and margin stays near the high 40s. The adverse scenario assumes shortages continue, memory costs force price increases, unit demand slows and Services remains below expectations.

The July 31 prices indicate that the market assigned more weight to Amazon’s favorable scenario and more weight to Apple’s near-term constraints. That is a snapshot, not a permanent verdict.

Principal Risks for Amazon

Capital expenditure outruns demand

Amazon is investing before all associated revenue is recognized. If cloud growth normalizes faster than expected, new data centers may operate below target utilization. The resulting depreciation, energy and maintenance costs would pressure margins while free cash flow remained weak.

Customer concentration in frontier AI

Large model developers can consume enormous capacity. That creates growth but may concentrate demand among a small number of customers with bargaining power and uncertain long-term economics. A funding slowdown, model consolidation or customer migration could affect usage.

Technology obsolescence

AI hardware evolves rapidly. Assets ordered today may face superior architectures before they reach full utilization. Custom chips reduce dependence on vendors but place design and software-execution risk on Amazon.

Anthropic investment volatility

Amazon’s investment relationship produced a large accounting gain in Q2. Future observable price changes can create gains or losses. The commercial and financing arrangements also add complexity that ordinary revenue growth does not capture.

Regulation and antitrust

Amazon operates a marketplace, logistics network, advertising platform and cloud provider. Regulators can scrutinize seller treatment, data use, competition, labor practices and acquisitions. Remedies may affect operating flexibility.

Retail cost pressure

Fuel, shipping, wages, tariffs and inventory placement can affect retail margin. Fast delivery improves demand but requires continuous network investment. International expansion adds currency and regional economic exposure.

Energy and supply constraints

Data centers require power, water, land, networking and advanced components. Delays in grid connections or shortages of memory and accelerators can slow deployment and raise costs.

Principal Risks for Apple

Supply constraints persist into key product launches

Apple’s September quarter typically includes important product transitions. Limited advanced chipmaking or memory supply can restrict availability precisely when demand and marketing activity are highest.

Price increases weaken unit demand

Apple can use price to offset higher component costs, but even loyal consumers have budgets. A broad increase may lengthen replacement cycles or shift demand to lower-priced models.

Services growth disappoints

Services supports the recurring-revenue and margin story. Continued results below consensus would reduce confidence that installed-base growth is translating into expected monetization.

Regulatory changes affect ecosystem economics

App distribution, payment rules, default services and platform access are under scrutiny. Changes can reduce commissions, increase competition or require costly product modifications.

AI features fail to stimulate upgrades

Apple’s AI strategy depends partly on improving the device experience. If features are delayed, undifferentiated or available on older hardware, the upgrade benefit may be smaller than expected.

China and geopolitical exposure

Greater China remains an important sales region and a critical part of Apple’s supply chain. Trade restrictions, competition, regulation and geopolitical tension can affect both demand and production.

A Practical Framework for Reading the Next Amazon Report

The next Amazon earnings release should be read in a sequence that prevents headline numbers from obscuring the economics.

1. Start with AWS growth. Compare the year-over-year rate with 37% and examine whether the acceleration continues, stabilizes or reverses.
2. Check AWS operating margin. Revenue growth is less valuable if infrastructure expense causes a sharp margin decline.
3. Review capital expenditure and free cash flow. Determine whether the $220 billion annual plan remains intact and whether operating cash flow keeps pace.
4. Separate operating income from net income. Identify investment valuation gains or losses before interpreting EPS.
5. Adjust retail growth for Prime Day timing. The reported Q3 comparison will not be directly comparable with Q2.
6. Examine advertising growth. Continued growth above the company average would support retail segment margins.
7. Read backlog and capacity commentary carefully. Reserved capacity is valuable only if contracts convert into profitable revenue.
8. Monitor depreciation and useful-life assumptions. The income-statement cost of the infrastructure build will rise over time.

This framework is more reliable than reacting to whether consolidated revenue beats consensus by one or two percentage points. Amazon is now a portfolio of businesses with radically different margins and capital needs. The mix determines the quality of the result.

A Practical Framework for Reading the Next Apple Report

1. Measure shipped demand, not only stated demand. Strong customer interest matters financially only when Apple can deliver products.
2. Compare gross margin with the 47% to 48% guide. Identify how much pressure comes from memory, product mix and tariffs.
3. Track Services against both growth and expectations. A record alone may not satisfy a premium valuation.
4. Watch iPhone pricing and unit indicators. Higher revenue can result from price, volume or mix, and each has different implications.
5. Review Mac performance after price changes. The June quarter benefited from strong demand, but future elasticity remains uncertain.
6. Examine Greater China. Growth was strong in Q3 but below one reported consensus figure; competitive and geopolitical conditions remain important.
7. Assess AI product adoption. The key question is whether new features improve upgrades, engagement or Services monetization.
8. Separate temporary benefits. Tariff refunds improved Q3 margin and EPS and should not be assumed to recur.

What Could Change the Amazon Thesis Over the Next Year

Four developments would strengthen the case. First, AWS growth remains above 30% even as the comparison becomes more difficult. Second, AWS margin stays in the upper 30s while depreciation rises. Third, advertising continues growing faster than consolidated revenue. Fourth, free cash flow begins recovering without a collapse in infrastructure investment.

A fifth positive would be clearer disclosure about AI and chip revenue. Management’s run-rate statements are useful, but investors would benefit from consistent definitions and trend data. Greater transparency would make it easier to distinguish broad enterprise adoption from a few large contracts.

The thesis would weaken if AWS growth falls sharply while capital expenditure remains elevated. A negative free-cash-flow period can be justified by acceleration; it is harder to justify during deceleration. Another warning would be a material decline in AWS margin without evidence that the reduction is buying durable share.

Changes in the Anthropic relationship also matter. An IPO or funding event could create further valuation movements and convert notes or preferred shares. Those events may be economically important but should remain separate from the assessment of operating performance.

Retail execution is the stabilizer. If North America and International margins remain positive, Amazon has more internal capacity to fund infrastructure. If retail costs rise at the same time AWS returns weaken, the consolidated risk increases.

What Could Change the Apple Thesis Over the Next Year

Apple’s near-term question is whether supply constraints are temporary. Improved advanced-node and memory availability would allow the company to convert demand, reduce emergency sourcing and protect product-launch execution. Continued constraints would make the September-quarter guide only the first sign of a longer problem.

Services is the second variable. Growth returning above expectations would reinforce the installed-base thesis. Repeated misses would suggest either more difficult comparisons, regulatory pressure or weaker monetization than the valuation assumes.

Pricing is the third variable. Apple has considerable brand power, but the memory cycle may force difficult choices. If customers accept higher prices without a material unit decline, revenue and margin can remain strong. If replacement cycles lengthen, the current product momentum may fade.

AI adoption is the longer-term variable. Apple does not need to sell raw computing capacity to benefit. It needs AI features that make devices more useful, encourage upgrades and strengthen ecosystem engagement. The financial evidence may appear gradually rather than in a single quarter.

Capital returns can provide support but cannot replace growth indefinitely. Apple’s dividend and repurchases reduce share count and return cash to investors. Their effect is most valuable when the operating business continues expanding and the repurchase price is reasonable.

How AWS Reached Its Fastest Growth Rate in 18 Quarters

AWS’s 37% growth is more meaningful when placed against the preceding six-quarter sequence disclosed by Amazon. Revenue growth was 17% in the first quarter of 2025, 17% in the second, 20% in the third, 24% in the fourth, 28% in the first quarter of 2026 and 37% in the latest quarter. The pattern is not a one-quarter bounce from a sudden collapse. It is a sustained acceleration across five consecutive steps after a period of slower expansion.

The revenue base increased at the same time. Quarterly AWS sales moved from $29.3 billion in Q1 2025 to $30.9 billion, $33.0 billion, $35.6 billion, $37.6 billion and then $42.2 billion. As a cloud business becomes larger, maintaining the same percentage growth requires more absolute revenue. The latest increase added roughly .4 billion from the year-earlier quarter. That is comparable to creating a substantial software company inside AWS in one year.

Operating-income history was less linear but equally important. AWS operating income was $11.5 billion in Q1 2025, $10.2 billion in Q2, $11.4 billion in Q3, $12.5 billion in Q4, $14.2 billion in Q1 2026 and $16.6 billion in Q2. The margin fell to 32.9% in Q2 2025 and then recovered through 34.6%, 35.0%, 37.7% and 39.4%. The sequence suggests improving utilization and revenue mix, although quarterly margin can also be affected by compensation, depreciation timing and foreign exchange.

Several forces can produce this type of curve. New data-center capacity may have become available after a constrained period. AI projects that began as tests may have moved into production. Customers may have committed to larger multi-year contracts. Amazon’s custom chips may have expanded the amount of compute it could offer at a competitive cost. Traditional cloud migrations may also have resumed after customers optimized spending during the earlier period of high interest rates and economic uncertainty.

The acceleration must not be extrapolated mechanically. Growth rates can peak when capacity catches up with an accumulated order book. Year-over-year comparisons become more difficult after an unusually strong quarter. Large contracts can create timing effects. AWS could remain healthy even if growth falls from 37% to the high 20s. The critical question is whether it settles at a structurally higher level than the 17% pace reported a year earlier.

There is another reason not to treat the 37% figure as a pure AI statistic. AWS sells thousands of services, and customers use its infrastructure for ordinary enterprise workloads as well as generative AI. Amazon does not disclose a complete split. The result demonstrates strong cloud consumption and supports the AI thesis, but it does not reveal exactly how much revenue came from training models, running inference, storing data or migrating conventional applications.

For investors and competitors, the sequence is nevertheless powerful. A cloud platform that can accelerate from 17% to 37% while expanding margin is gaining economic momentum. The burden of proof now shifts to sustainability. Future disclosures must show whether the acceleration reflects a durable adoption wave or a temporary release of constrained demand.

From Operating Cash Flow to Negative Free Cash Flow: The Mechanics

Amazon’s cash-flow statement can appear contradictory. Operating cash flow increased by 33% to $161.4 billion over the trailing 12 months, yet free cash flow fell below zero. Both statements are true because they measure different stages of the cash cycle.

Operating cash flow begins with net income and adjusts for non-cash items and working-capital movements. Amazon’s large Anthropic valuation gain increased net income but was non-cash, so it is removed in the operating-cash-flow reconciliation. Depreciation is added back because it reduces accounting profit without using cash in the current period. Changes in payables, receivables, inventories and other working-capital accounts also affect the total.

Free cash flow, under Amazon’s definition, subtracts purchases of property and equipment, net of proceeds and incentives, from operating cash flow. Those net purchases reached $169.0 billion over the trailing year, exceeding operating cash flow by about $7.6 billion. The result is negative free cash flow even though the underlying operations generated a very large amount of cash.

This distinction is essential because capital expenditure can create productive assets. A retailer’s inventory purchase is usually a working-capital item that turns into cost of sales when goods are sold. A data center, server or network asset is capitalized and used over several years. The cash outflow occurs at purchase, while depreciation is spread through future income statements.

The current period therefore contains an intertemporal trade. Amazon accepts lower cash generation now in exchange for expected cloud and logistics capacity later. The trade creates value only if future after-tax cash inflows exceed the cost of building, financing and maintaining the assets.

Investors should monitor more than the top-line free-cash-flow number. Cash capital expenditures reached $53.1 billion in Q2 alone. Technology and infrastructure operating expense rose 22%, reflecting depreciation and other costs already entering the income statement. Interest expense increased to $1.3 billion from $516 million, partly because of new debt and lease obligations. Amazon said it expects additional financing activity in 2026.

These trends show how a capital cycle moves through the financial statements. First comes cash spending. Then assets enter service and depreciation rises. Financing can add interest expense. Revenue and operating profit must grow fast enough to cover all three. Q2’s AWS result passed the current test, but the installed asset base will continue generating costs even if demand later slows.

Working capital can also create volatility. Amazon collects cash from consumers quickly and pays many suppliers later, which can make growth a source of cash. Prime Day timing, inventory build and payable schedules can shift operating cash flow among quarters. Trailing-12-month figures reduce some seasonality but do not eliminate it.

A useful forward indicator is capital expenditure as a share of operating cash flow. On the trailing figures, net property-and-equipment purchases exceeded operating cash flow. That ratio is unlikely to remain sustainable indefinitely without additional financing or a future increase in cash generation. Management’s argument is that 2026 represents a build phase whose assets will support higher revenue in later years.

The skeptical case is that AI infrastructure may require continuing replacement and expansion, making the “temporary” build more permanent. Accelerators can have shorter economic lives than conventional buildings, and model demand can rise as fast as efficiency improves. If every generation requires another extraordinary investment round, free cash flow may not rebound as quickly as a traditional capacity cycle would suggest.

The favorable case is that scale, custom silicon and high utilization improve returns. Once power, land and buildings are in place, incremental hardware and software can expand output. Reserved capacity reduces demand risk. High-margin services layered on the infrastructure can raise revenue per dollar of assets. The next two years will reveal which effect dominates.

Amazon’s Investment Portfolio Has Become Material to Reported Earnings

The Anthropic disclosure highlights a broader change in Amazon’s financial profile. The company is not only a buyer of infrastructure and a provider of cloud services; it is also a large strategic investor in model developers. Those investments can create commercial alignment and substantial accounting volatility.

Amazon’s 10-Q says it invested in Anthropic preferred stock, amended commercial arrangements involving AWS services and chip-performance obligations, and established a financing structure linked to compute-capacity milestones and a future liquidity event. It also disclosed large investments in OpenAI during the first half of 2026. These relationships tie capital, cloud contracts and strategic positioning together.

From a business perspective, the logic is clear. Model developers need compute. Cloud providers need anchor customers and workloads that test their hardware. Equity ownership gives Amazon exposure to the economic value of successful model companies. Contractual chip use can accelerate adoption of Trainium. Bedrock can distribute models to enterprise customers.

From an accounting perspective, the relationships complicate earnings analysis. Observable funding prices can lead to valuation adjustments. Convertible instruments can move between available-for-sale debt securities and equity classifications. Unrealized gains can appear in other comprehensive income or net income depending on the instrument and event. A future IPO could trigger conversions and additional remeasurement.

These are not merely technical footnotes when the gain reaches tens of billions of dollars. They can dominate quarterly net income and distort comparisons with analyst estimates based on recurring operations. They can also affect taxes, balance-sheet assets and future volatility.

The economic exposure is two-sided. A higher Anthropic valuation increases Amazon’s reported wealth and may validate the strategic relationship. A lower valuation could produce losses. If the model developer succeeds commercially but relies on expensive AWS commitments, the value may circulate between investment returns and cloud revenue. If it fails to generate sustainable cash flow, both the investment and associated demand assumptions can be questioned.

Readers should therefore separate three questions. Is AWS attracting and retaining profitable cloud demand? Are Amazon’s strategic investments appreciating? Are the contractual relationships creating concentration or financing risk? A positive answer to one does not guarantee a positive answer to the others.

More detailed disclosure would help. Investors would benefit from clearer information about the duration and economics of major capacity commitments, the percentage of AWS growth attributable to a few model developers, and the sensitivity of investment carrying values. Commercial confidentiality limits what Amazon can reveal, but the materiality of the figures makes transparency increasingly important.

Scenario Analysis: What the Next Four Quarters Could Look Like

Forecasting a company with Amazon’s scale requires ranges rather than a single target. The following scenarios are analytical frameworks, not company guidance or investment recommendations.

Scenario One: AWS remains above 30% growth

In the constructive scenario, AI demand, cloud migrations and reserved capacity keep AWS growth above 30% through most of the next year. AWS margin remains in the upper 30s as custom chips and better utilization offset depreciation. Advertising continues growing above 20%, and retail margins remain positive. Capital expenditure stays high in 2026 but begins growing more slowly in 2027. Operating cash flow catches up, and free cash flow moves back toward positive territory.

This scenario would support the market’s July repricing. It would show that Q2 was an inflection rather than a peak. The main risk within the constructive case is valuation: even excellent execution may already be partly reflected in a higher share price.

Scenario Two: growth normalizes but profitability holds

In a middle scenario, AWS growth falls into the mid-to-high 20s as comparisons become harder. Segment margin remains around the mid-to-upper 30s. Retail and advertising deliver steady growth. Capital expenditure remains near the current plan, keeping free cash flow around break-even or modestly negative.

This would still represent a strong business, but the stock response would depend on expectations. A decline from 37% to 27% can look disappointing even when the absolute revenue addition remains enormous. Investors would focus on whether the slowdown reflected capacity timing, normal comparison effects or lost competitive momentum.

Scenario Three: capital pressure arrives before revenue

In the adverse scenario, AWS growth drops below the mid-20s while depreciation and energy expense accelerate. Large customers delay deployments, optimize workloads or negotiate lower prices. Amazon continues spending because projects are already committed. AWS margin falls, and free cash flow remains materially negative.

This scenario would revive concerns that the AI build exceeded near-term demand. It could also pressure suppliers if Amazon reduced future orders. The company’s retail and advertising profits would provide a cushion, but they would not fully offset a large cloud-margin decline.

Scenario Four: supply, not demand, limits AWS

A separate scenario is constrained growth. Customer demand remains very strong, but power, networking, memory or accelerator shortages prevent Amazon from deploying capacity quickly enough. Reported AWS growth slows even as backlog rises. Capital expenditure remains high, and management emphasizes future reservations.

This case is difficult to value because weak current revenue coexists with strong claimed demand. Investors would need contract evidence, delivery timelines and signs that constraints are resolving. The experience of Apple’s quarter shows that markets can punish supply-limited guidance even when management says demand is healthy.

The most useful signal across all scenarios is the joint movement of AWS revenue, AWS operating income and cash capital expenditure. Revenue alone cannot show returns. Margin alone can be helped by timing. Capex alone cannot show demand. Together, the three reveal whether Amazon is converting investment into profitable growth.

Scenario Analysis for Apple’s Next Two Product Cycles

Supply normalizes and demand remains strong

Apple’s favorable scenario begins with improved access to advanced manufacturing and memory. The company fulfills more iPhone and Mac demand, avoids excessive component premiums and holds gross margin in the high 40s. Services returns to growth above expectations, and new AI features support upgrades without requiring disruptive price increases.

This would make the July selloff look primarily like a timing reaction. Deferred demand could move into later quarters, and stronger availability could restore revenue growth. Apple’s installed base and cash generation would remain central strengths.

Demand is strong but pricing does more of the work

In a mixed scenario, component costs stay high and Apple raises prices. Revenue grows, but unit volumes become less certain. Premium models perform well while lower-income customers extend replacement cycles. Gross margin is protected, but the addressable upgrade pool narrows.

This outcome can produce good earnings in the short term and weaker ecosystem growth later. It would be important to distinguish average selling price from unit demand, though Apple does not disclose product units.

Supply constraints and Services softness persist

In the adverse scenario, Apple cannot secure enough advanced chips or memory, gross margin falls toward the low end of guidance, and Services repeatedly misses expectations. Consumers delay upgrades after price increases, and regulatory changes affect App Store or distribution economics.

Apple would remain profitable, but the valuation premium could compress because both hardware execution and recurring-revenue growth would be questioned. The company’s repurchases could soften the per-share effect but would not solve an operating slowdown.

AI features create a new upgrade cycle

A separate upside scenario is that new Siri and on-device intelligence features become compelling enough to stimulate upgrades across iPhone, Mac and other devices. The benefit would appear in product mix, units, Services engagement and possibly new paid offerings. Apple would not need to disclose a stand-alone AI revenue line for the economic impact to become visible.

The challenge is causality. Product cycles, pricing, carrier promotions and macroeconomic conditions also affect sales. Investors should look for persistent improvements rather than attributing every strong quarter to AI.

What the Amazon–Apple Split Says About Earnings Journalism

The immediate market narrative reduced the reports to a simple contrast: Amazon surged and Apple fell. That is useful for describing price action but insufficient for understanding the businesses.

Amazon’s headline EPS was the least reliable measure of recurring performance because of the Anthropic valuation gain. Its negative free cash flow was the clearest financial cost of the investment strategy. Its operating income and AWS segment results were the strongest evidence in favor of that strategy. All four facts belong in the same assessment.

Apple’s share decline did not mean its reported quarter was poor. Revenue, EPS, iPhone and Mac sales all grew strongly. The weakness lay in the difference between reported performance and forward expectations. Tariff refunds also improved the quarter, making normalization important.

Good earnings analysis therefore separates reported facts, consensus comparisons, accounting effects, management explanations and market interpretation. It also timestamps prices. An after-hours move can differ from the next day’s close; an early-session percentage can differ from the final result. Amazon’s initial after-hours gain and July 31 close were both positive, but the final 15.25% move provides the complete first-session reaction.

Management statements should be treated as claims supported or challenged by financial data. Amazon says it is building for existing demand. Reserved capacity and AWS growth support that claim, while negative free cash flow shows the cost. Apple says supply, not demand, limits the outlook. Strong product revenue supports that claim, while the lower guide shows the consequence.

The most credible conclusion is rarely found in one number. Amazon’s quarter was operationally strong and financially cash-intensive. Apple’s quarter was historically strong and forward-lookingly constrained. The market reaction emphasized one side of each statement, but long-term outcomes depend on both.

Frequently Asked Questions

Did Amazon beat earnings expectations in Q2 2026?

Amazon reported stronger-than-expected sales and AWS growth, according to market estimates cited by major news organizations. The most important operational beat was AWS revenue of .2 billion and 37% growth. Headline EPS of $5.75 was heavily affected by a large non-operating valuation gain related mainly to Anthropic, so it should not be interpreted as a purely recurring earnings beat.

How much revenue did Amazon report?

Amazon reported 0.6 billion of net sales for the quarter ended June 30, 2026, up 20% from 7.7 billion a year earlier.

How fast did AWS grow?

AWS revenue increased 37% year over year to .2 billion. Amazon said this was the cloud unit’s fastest growth in 18 quarters and equivalent to an annualized revenue run rate of approximately 9 billion.

How profitable was AWS?

AWS produced $16.6 billion of operating income, up 64% from $10.2 billion. Its operating margin was 39.4%, compared with 32.9% in the year-earlier quarter.

Why was Amazon’s net income so high?

Amazon recorded $53.4 billion of non-operating income, driven primarily by upward valuation adjustments and reclassification gains connected with Anthropic investments. That helped lift net income to $62.6 billion. Operating income of $27.5 billion is a cleaner measure of the quarter’s core business performance.

Why was Amazon free cash flow negative?

Trailing-12-month operating cash flow reached $161.4 billion, but net property-and-equipment purchases rose to $169.0 billion, primarily because of AI infrastructure investment. Amazon’s resulting free cash flow was an outflow of $7.6 billion.

How much does Amazon plan to spend on capital expenditures in 2026?

Management said it expects approximately 0 billion of capital expenditures in 2026, up from an earlier 0 billion plan. The spending includes technology infrastructure as well as robotics, chips, satellites and fulfillment capacity, with the majority of technology infrastructure directed toward AWS.

Why did Amazon stock rise after earnings?

Investors responded to the combination of 37% AWS growth, a 39.4% AWS operating margin, higher operating income and management’s comments about reserved future cloud capacity. Amazon closed July 31, 2026, at $271.58, up 15.25%.

Did Apple have a bad quarter?

No. Apple reported record June-quarter revenue and EPS, with total revenue up 16% to $109.4 billion and diluted EPS up 29% to $2.02. The stock fell because Services revenue missed expectations, part of the earnings benefit came from tariff refunds, and the September-quarter outlook was below Wall Street’s target.

Why did Apple stock fall?

Apple forecast 9% to 11% revenue growth for the September quarter, below the approximately 12% consensus reported by Reuters. Management also described significant supply constraints, while Services revenue of $30.7 billion fell short of the expected level. Apple shares closed down 7.14% on July 31.

How much did Apple earn per share?

Apple reported diluted EPS of $2.02. The company said tariff refunds contributed $0.11. Excluding that benefit, EPS would have been approximately $1.91, still above the $1.89 estimate cited by Reuters.

What should readers watch next?

For Amazon, the priorities are AWS growth, AWS margin, capital expenditure, free cash flow and the conversion of reserved capacity into revenue. For Apple, the priorities are component availability, September-quarter revenue, gross margin, Services growth, pricing and the financial impact of new AI features.

Final Assessment

Amazon’s second quarter provided the strongest current evidence that its AI infrastructure spending is producing measurable commercial returns. AWS did not merely grow; it reaccelerated to 37%, generated $16.6 billion of operating income and expanded its margin to 39.4%. Retail, marketplace services and advertising also contributed, lifting consolidated operating income 43% to $27.5 billion.

The result deserves a more careful reading than the $5.75 EPS headline. Amazon’s net income was enlarged by a major Anthropic-related valuation gain. The operating result is compelling without it, which is precisely why the distinction strengthens rather than weakens the quarter’s central message. Amazon’s core businesses performed well.

The credible concern is cash intensity. A $220 billion capital-expenditure plan and negative trailing free cash flow create a high requirement for future utilization. Amazon must keep filling data centers, converting reservations and defending margins while equipment, power and depreciation costs rise. The company earned greater investor tolerance for spending in Q2; it did not earn permanent immunity from capital discipline.

Apple’s quarter offered the mirror image. Its completed results were excellent: record June-quarter revenue, strong iPhone and Mac sales, high margin and nearly $30 billion of net income. The market focused on what Apple could not guarantee—enough supply, enough Services upside and enough September-quarter growth to satisfy elevated expectations.

The earnings split therefore says less about one company being “good” and the other “bad” than about the standard applied to megacap technology in 2026. Investors rewarded a company whose spending produced immediately visible cloud growth and punished a company whose strong demand encountered operational constraints. The next phase will test whether Amazon’s capacity build remains productive and whether Apple can turn product enthusiasm into delivered units and recurring services revenue.

For Amazon, the decisive evidence will be the relationship between AWS growth and free cash flow. For Apple, it will be the relationship between demand and supply. Both companies remain highly profitable. Their risks now arise from the scale of the systems required to sustain that profitability.

Sources

• Amazon.com Announces Second Quarter Results
• Amazon Form 10-Q for the quarter ended June 30, 2026
• Amazon.com Announces First Quarter 2026 Results
• Apple Reports Third Quarter Results
• Apple fiscal third-quarter 2026 earnings release filed with the SEC
• Reuters: Amazon soars as cloud revenue surge allays fears over ballooning AI bets
• Reuters: Apple disappoints with forecast dogged by supply chain struggles
• Reuters: Wall Street ends higher as Amazon soothes AI jitters
• Associated Press: Amazon increases technology investment plan after strong cloud growth

Affiliate disclosure: Businessfinance.news may earn compensation from qualifying actions completed through selected links on this website, at no additional cost to the reader. Affiliate relationships do not influence our editorial reporting, analysis, or conclusions.