Amazon Surges 15% as Big Tech’s AI Spending Faces a Profit Test

Amazon shares jumped more than 10% before Friday’s opening bell after AWS delivered its fastest growth in 18 quarters. The move placed Amazon on the winning side of an increasingly sharp divide across Big Tech.

A premarket snapshot showed Amazon leading most large technology stocks higher. Intel gained more than 4%, while Alphabet, Nvidia, Meta, and Tesla posted smaller advances. Apple fell almost 8%, and Microsoft slipped slightly.

By the closing bell on July 31, Amazon had climbed 15.3%, while Apple had lost 7.4%. The contrast was not simply a reaction to two earnings reports. It showed how investors are evaluating the enormous cost of building AI infrastructure.

Wall Street is no longer rewarding every company that promises to spend heavily on artificial intelligence. It is rewarding companies that connect those investments to accelerating revenue and operating profit.

Amazon cleared that test through AWS. Apple failed a different test because its outlook exposed the pressure that AI infrastructure demand is placing on memory supplies and hardware costs. Meta remained between those extremes after reporting fast revenue growth alongside an even faster increase in expenses.

The result was a revealing market split. AI investment remains attractive, but spending alone has stopped functioning as a bullish signal.

Amazon’s Rally Started With AWS, Not Retail Optimism

Amazon’s 15.3% closing gain reflected evidence that cloud and AI investments were producing measurable operating results.

Amazon reported second-quarter net sales of $200.6 billion, an increase of 20% from the same period in 2025. Operating income rose 43% to $27.5 billion, according to the company’s quarterly results.

AWS supplied the most important numbers. The cloud division generated $42.2 billion in quarterly sales, up 37% year over year. Amazon said the 36.7% unrounded growth rate was AWS’s fastest in 18 quarters.

AWS operating income reached $16.6 billion, compared with $10.2 billion one year earlier. That increase gave investors something more convincing than a broad statement about future AI demand.

Amazon also said its AWS AI business and custom chip business each exceeded a $25 billion annual revenue run rate. A run rate projects recent revenue over a full year, rather than representing completed annual sales.

Those company-reported figures have not been independently audited as separate business segments. However, they offer a clearer commercial signal than customer counts or model announcements alone.

The cloud acceleration mattered because Amazon spent the previous several quarters defending its position against Microsoft Azure and Google Cloud. Both rivals had reported faster percentage growth while promoting their relationships with leading model developers.

AWS still had enormous scale, but investors wanted proof that generative AI demand would accelerate its growth rate. The latest quarter supplied that proof.

The underlying mechanism extends beyond renting Nvidia processors. Amazon has invested in Trainium AI accelerators, Graviton server processors, Bedrock model services, and data centers designed around those products.

Amazon said Anthropic and OpenAI had made multiyear, multigigawatt commitments involving Trainium. The company also reported increasing adoption among startups and larger customers, including Uber and Pinterest.

These claims support Amazon’s argument that owning more of the infrastructure stack can improve both supply availability and economics. Custom chips can reduce dependence on outside suppliers, while AWS services create additional opportunities to capture customer spending.

The quarter did not prove that Amazon has won the AI cloud race. It demonstrated that its infrastructure spending was accompanied by faster revenue growth and higher operating income.

That distinction explains why Amazon’s stock response exceeded the smaller premarket gains recorded by Nvidia, Alphabet, and Meta. Investors were reacting to an earnings conversion, not merely another AI product cycle.

Amazon’s noncloud businesses also contributed. North American sales rose 16% to $116.2 billion, while international sales increased 15% to $42.2 billion.

Advertising revenue grew 26%, according to the company. Faster delivery and expanding everyday-goods sales also strengthened the retail operation.

Still, those results did not create the central market narrative. AWS transformed a strong quarter into a referendum on whether Amazon’s AI capital spending was beginning to pay back.

Why Wall Street Treated Amazon and Apple So Differently

Amazon and Apple both reported strong quarters, but their outlooks pointed toward opposite sides of the AI infrastructure cycle.

Apple earned $29.79 billion during its fiscal third quarter, up 27% from the prior year. Revenue increased 16% to $109.42 billion, according to its reported fiscal results.

Those figures exceeded average analyst expectations tracked by FactSet. iPhone, Mac, and Services revenue grew at double-digit rates, while every reported geographic segment expanded.

Ordinarily, that combination would support the stock. Instead, Apple shares dropped because its forecast for the current quarter fell short of Wall Street’s expectations.

Executives attributed part of the pressure to component shortages. AI data-center construction has increased demand for memory chips, creating competition between infrastructure buyers and consumer-device manufacturers.

That supply relationship turned the Amazon and Apple moves into two sides of the same trade. Amazon benefited as companies purchased more cloud capacity. Apple faced higher costs and tighter supplies for devices sold to consumers.

The contrast also challenged a familiar assumption about Apple. The company had previously appeared insulated from the spending race consuming cash at Amazon, Meta, Microsoft, and Alphabet.

Apple did not need to match hyperscaler capital expenditures to maintain its consumer ecosystem. Hyperscalers operate massive data centers and sell computing capacity to outside customers.

However, Apple still depends on the same semiconductor supply chain. It can avoid building as many AI data centers, but it cannot avoid the effect of infrastructure demand on memory availability and component costs.

Apple had already raised prices for some Macs and iPads, citing the memory shortage. It had not announced a comparable iPhone increase when it reported earnings.

Investors therefore looked beyond Apple’s completed quarter. They focused on whether stronger component costs would weaken margins, constrain production, or require additional consumer price increases.

Amazon presented the opposite timeline. Its heavy spending had already reduced cash generation, but AWS growth and operating income showed a visible return.

The market’s judgment was not that Amazon had lower risk. Amazon carries substantial execution risk because it must build capacity before all demand becomes certain.

Instead, investors judged Amazon’s current evidence as stronger. Its fastest AWS expansion in 18 quarters created a direct line from infrastructure investment to revenue growth.

Apple’s earnings beat could not erase a weaker outlook. In a market focused on future AI economics, guidance mattered more than the quarter already completed.

The pattern extended beyond these two companies. Microsoft had rallied one day earlier after showing that Azure growth and profit could absorb rising infrastructure costs.

Meta received a more skeptical reaction because its expense growth outpaced its revenue growth. The companies were reporting different businesses, but Wall Street applied the same test.

Investors wanted to know whether spending was producing near-term operating leverage. Operating leverage occurs when revenue grows faster than major costs, allowing profit to expand more quickly.

Amazon and Microsoft supplied that pattern. Meta did not, while Apple faced an indirect cost shock from the same investment boom.

Meta Shows the Other Side of the AI Spending Race

Meta’s results showed why strong revenue growth no longer guarantees a positive response when AI expenses rise even faster.

Meta reported second-quarter revenue of $60.8 billion, an increase of 28% year over year. Its advertising business remained healthy across Facebook, Instagram, and its broader application portfolio.

However, expenses increased 55% to roughly $42 billion. Net income declined 14% to $15.8 billion, according to an earnings analysis.

Meta also projected 2026 capital expenditures between $130 billion and $145 billion. The company raised the lower end of that range while leaving the upper end unchanged.

Capital expenditures include purchases of long-lived assets such as servers, networking equipment, and data-center facilities. They do not immediately appear as operating expenses, but they consume cash and create future depreciation charges.

Meta’s position differs from Amazon’s because it does not operate a comparable public cloud business. Its AI systems primarily improve advertising, recommendations, content creation, and consumer products.

That model can produce substantial returns. Better recommendations can increase user engagement, while better advertising tools can improve conversion rates and campaign performance.

The challenge is attribution. Amazon can report AWS revenue and operating income as separate figures. Meta cannot isolate an equally clean revenue stream generated by each new data center or model.

Investors must infer the return through advertising growth, engagement, and future products. That leaves more room for disagreement about the timing and durability of the payoff.

Meta’s smaller premarket gain on Friday followed a 6.2% decline in after-hours trading after its results. The shifting reaction reflected uncertainty rather than a stable endorsement.

The company is still growing faster than many businesses of comparable size. Yet its expense increase showed that the AI race can weaken earnings even when the core advertising engine remains strong.

This is the main pressure facing Meta. It must continue investing because competitors are improving models, recommendation systems, and advertising automation.

Reducing spending could protect near-term cash flow but weaken Meta’s future position. Maintaining the current pace requires convincing investors that revenue and profit will eventually catch up.

Microsoft faces a related challenge, although Azure gives it a direct channel for selling AI capacity. The company said capital expenditures rose 70% to $41 billion during its latest quarter.

Microsoft also reported a 31% increase in net income to $35.8 billion. That combination helped investors tolerate the higher spending.

The difference was operating leverage. Microsoft’s profit increased despite the infrastructure bill, while Meta’s profit declined.

Amazon’s rally therefore increased the pressure on Meta to provide measurable evidence. General claims about engagement or long-term superintelligence ambitions will carry less weight after AWS delivered visible acceleration.

The comparison does not mean Amazon’s strategy is inherently better. Amazon, Microsoft, and Meta monetize AI infrastructure through different products and accounting structures.

It does mean investors are building a hierarchy of evidence. Segment revenue and operating profit sit near the top. Engagement claims and future product promises sit lower until they appear in financial results.

For enterprise buyers, this split has practical implications. Aggressive spending can improve model availability, custom chips, agent services, and cloud capacity.

It can also create pressure to monetize those investments through larger customer commitments, consumption growth, or broader product bundles. Buyers should expect hyperscalers to connect more services to their AI platforms.

The Rally Did Not Remove Amazon’s Cash Flow Risk

Amazon’s earnings strengthened the AI investment case, but the quarter also exposed how much cash that strategy requires.

Amazon reported a trailing 12-month free cash outflow of $7.6 billion. One year earlier, the company had generated $18.2 billion in free cash flow.

Free cash flow measures cash remaining after capital investments. It helps show how much financial flexibility a company retains after maintaining and expanding its operations.

Amazon attributed the reversal primarily to a $66.1 billion year-over-year increase in property and equipment purchases. The company said those purchases mainly reflected AI investments.

That cash flow decline is the strongest counterargument to the bullish market response. AWS growth accelerated, but Amazon is paying heavily to support that growth.

Net income also requires careful interpretation. Amazon reported $62.6 billion in quarterly net income, more than triple the prior-year figure.

However, the total included $53.4 billion in pretax nonoperating income, primarily connected to Amazon’s investment in Anthropic. That gain did not come from ordinary retail or AWS operations.

Operating income provides a cleaner view of the underlying businesses. It still increased 43% to $27.5 billion, which was a strong result without the investment gain.

The distinction matters because headline profit can make the quarter appear less risky than it was. Amazon’s core operations improved, but cash generation moved in the opposite direction.

Data-center construction also creates a timing mismatch. Companies commit capital before servers begin producing revenue, and capacity can become less valuable if demand changes.

AI hardware has shorter economic cycles than many traditional infrastructure assets. New processors can deliver better performance or efficiency, reducing the attractiveness of older equipment.

Microsoft highlighted a similar issue by saying much of its capital spending involved shorter-lived assets, primarily CPUs and GPUs. Those assets require replacement as computing systems advance.

Amazon has some protection because it operates at enormous scale and designs custom chips. It can allocate capacity across many customers and services.

Its customer commitments may also improve visibility. Amazon said leading AI companies had made multiyear infrastructure commitments, suggesting that some planned capacity already has contracted demand.

Still, contracted demand does not eliminate execution risk. Construction delays, power constraints, chip supply, pricing pressure, or weaker customer usage can change the expected return.

The company’s third-quarter guidance offered another checkpoint. Amazon projected net sales between $197 billion and $202 billion, representing growth of 9% to 12%.

It projected operating income between $22.5 billion and $26.5 billion. Both ranges remain subject to currency movements, energy costs, trade policies, inflation, and customer spending.

The wider market environment adds another layer of uncertainty. The 10-year Treasury yield rose to 4.71% on Friday as investors worried about persistent inflation.

Higher yields can reduce the present value investors assign to future technology profits. They also raise financing costs for businesses throughout the economy.

Oil prices and geopolitical instability added to those concerns. Brent crude settled at $87.93 after large swings during July.

Despite those pressures, the broader market finished higher. The S&P 500 gained 0.7%, the Dow rose 0.5%, and the Nasdaq advanced 1%.

Amazon accounted for an important part of that optimism. However, a single earnings reaction does not settle whether the current investment cycle will produce durable returns.

The next test is repetition. AWS must sustain strong growth while Amazon limits further deterioration in free cash flow.

Amazon’s Gain Raises the Bar for Every Hyperscaler

The market is shifting from rewarding AI exposure to comparing the quality, timing, and visibility of AI returns.

For much of the generative AI cycle, companies received credit for securing GPUs, announcing data centers, or partnering with model developers. Those actions demonstrated access to scarce infrastructure.

Capacity is no longer enough. Amazon’s quarter gave investors a stronger benchmark built around growth acceleration, segment profit, and identifiable customer demand.

Microsoft already occupies a similar position. Azure provides direct AI infrastructure revenue, while productivity software creates another path for monetization.

Alphabet combines Google Cloud with advertising and consumer AI services. Its challenge resembles Microsoft’s, although search competition adds another source of pressure.

Meta lacks a public cloud segment, so it must show returns through advertising performance and consumer adoption. That makes expense discipline more important to its investment story.

Apple faces a different constraint. It controls a major consumer platform but relies on external infrastructure and component suppliers for parts of its AI strategy.

Nvidia remains an upstream beneficiary because it sells the processors used across many competing platforms. However, custom chips from Amazon, Google, and others create long-term competitive pressure.

Intel’s premarket gain of more than 4% showed that investors were willing to extend some optimism across the semiconductor sector. Yet the day’s later volatility illustrated the limits of that read-through.

Micron rose as much as 6.4% early in Friday’s session before closing down 5.9%. That reversal showed how quickly enthusiasm about AI demand can collide with valuation and supply-cycle concerns.

The market is therefore separating companies along three dimensions.

First, investors are comparing revenue visibility. AWS and Azure can directly bill customers for computing, storage, databases, and AI services.

Second, they are comparing margin conversion. Revenue growth carries more weight when operating income expands alongside it.

Third, they are comparing capital intensity. Similar growth rates can receive different valuations when one requires much more cash.

Amazon scored well on the first two dimensions this quarter. It remained exposed on the third.

Meta scored well on revenue growth but poorly on expense growth. Apple generated strong profit but presented a weaker outlook tied partly to AI-driven component scarcity.

These differences explain why “Big Tech rose” is an incomplete account of Friday’s trading. The sector did not move as a single group.

The premarket figures included Amazon above 10%, Meta up 0.89%, Nvidia up 0.9%, Tesla up 0.67%, and Alphabet above 1%. Those percentages looked broadly positive.

Yet their underlying stories were different. Amazon had just reported accelerating cloud growth. Meta was recovering from an earnings-related decline. Nvidia remained tied to infrastructure demand across the entire group.

Apple’s decline offered the clearest counterweight. Its strong completed quarter did not protect the stock from a disappointing forward view.

For technology executives, the same distinction should shape planning. AI spending deserves scrutiny based on measurable business outcomes, not the size of the budget.

Cloud commitments, model usage, developer adoption, and workflow automation can all support a return. None automatically guarantees profitable deployment.

Teams also need systems for comparing announcements with later evidence. A searchable knowledge workflow can help track financial claims, product releases, and customer adoption across reporting periods.

That discipline matters because AI vendors increasingly publish many overlapping performance measures. Run rates, customer counts, token volumes, and engagement rates answer different questions.

Investors rewarded Amazon because several measures aligned. AWS growth accelerated, AWS operating income increased, and the company described large customer commitments.

The remaining weakness was cash flow. That is why Amazon’s gain raised the bar without ending the debate.

Three Signals Will Decide Whether the Amazon Rally Holds

The next three months must show whether Amazon produced a repeatable shift or one unusually strong quarter.

The first signal is AWS growth. Investors should watch whether AWS maintains a growth rate near its latest 36.7% pace when Amazon reports again.

A second quarter of strong expansion would strengthen the argument that enterprise AI workloads are moving into sustained production. A sharp slowdown would make Friday’s rally look more dependent on temporary demand or easy comparisons.

The mix of that growth matters as much as the percentage. Customers experimenting with models can create bursts of computing demand without establishing durable applications.

Production deployments usually require databases, security controls, networking, monitoring, and long-term infrastructure commitments. Those adjacent services can make AI revenue more persistent.

Amazon’s custom chips will also influence the outcome. Broader Trainium adoption would support Amazon’s claim that it can compete through both infrastructure scale and hardware economics.

The second signal is free cash flow. AWS can continue growing rapidly while Amazon’s valuation comes under pressure if capital spending keeps consuming more cash.

Investors should compare equipment purchases, operating cash flow, and AWS operating income. Improvement across all three would show that growth is becoming more financially efficient.

Another large decline in free cash flow would weaken that conclusion, even if revenue remains strong. It would suggest that each new stage of growth still requires an outsized capital commitment.

The third signal is the response from Microsoft, Google, and Meta. These companies must now explain how their own spending translates into revenue, profit, or measurable customer outcomes.

Microsoft will need to maintain Azure growth while absorbing the depreciation associated with rapidly expanding infrastructure. Google must show that cloud momentum can coexist with pressure on its search business.

Meta faces the clearest near-term comparison. Its revenue grew 28%, but expenses increased 55% and net income declined.

If Meta produces stronger profit conversion during its next report, investors may treat the latest weakness as a temporary investment phase. Continued expense growth without matching earnings would strengthen Amazon’s relative position.

Apple’s next product cycle will offer a related signal from outside the hyperscaler group. Investors should watch component availability, device pricing, and management’s margin outlook.

Easing memory constraints would weaken the idea that AI infrastructure demand is materially pressuring consumer hardware. Persistent shortages would reinforce it.

These signals matter beyond stock traders. Developers and enterprise customers depend on the same investment cycle for computing availability, service reliability, and AI deployment costs.

A healthy cycle should produce more capacity, better chips, broader model access, and lower unit costs. An overheated cycle can produce shortages, aggressive commitments, and pressure to increase customer spending.

Amazon’s quarter supplied the strongest evidence yet that its recent AI investments are generating substantial operating returns. The company still has to show that those returns can outlast the construction bill.

The market has delivered its initial verdict. Amazon gained 15.3%, Apple fell 7.4%, and Meta remained under scrutiny despite strong revenue growth.

The next verdict will come from operating results rather than announcements. Watch AWS growth first, free cash flow second, and competitor profit conversion third.

For anyone tracking Big Tech’s AI race, that is the useful framework. Ask which company is spending, where the revenue appears, and how much cash remains after the infrastructure is built.