Amazon's Earnings Land Right in the Middle of Big Tech's AI Spending Freakout

Amazon reports second-quarter earnings after the bell on Thursday, walking into a week that has already scrambled how investors think about Big Tech's AI spending spree. The report will test whether Amazon can pull off the trick Microsoft just managed, and avoid the one that hit Meta.

Analysts polled by LSEG expect Amazon to post earnings of $1.82 per share on revenue of $196.47 billion. Within that, Wall Street will be watching two numbers closely: Amazon Web Services revenue, pegged at $40.54 billion by StreetAccount, and advertising revenue, expected to come in around $19.43 billion.

But the number that's really moving markets this earnings season isn't revenue. It's capital expenditures, the money companies pour into data centers, chips and other infrastructure to keep their AI ambitions running. That spending has turned into the central drama of this earnings cycle, and it's producing wildly different reactions depending on what comes with it.

A rough week to be a hyperscaler

Alphabet kicked things off by raising its 2026 capex forecast to as high as $205 billion, and its shares sank on the news. Then Microsoft reported stronger-than-expected earnings and reaffirmed its 2026 capex plans, and its stock jumped as much as 15%. Meta went the other direction: it gave investors a light revenue forecast for the current quarter, warned that its AI bets were eating into cash flow, and watched its stock drop 9%.

The pattern is clear enough. Investors aren't punishing big AI spending on principle. They're punishing spending that doesn't come with a convincing growth story attached.

That's the needle Amazon has to thread. The company's capital expenditures hit $44.2 billion in the first quarter, up 77% from a year earlier, and FactSet data suggests that figure climbed further to $49.3 billion in the second quarter. Amazon has so far stuck with its February guidance of roughly $200 billion in capex for 2026, but several analysts think it's about to follow Alphabet's lead and raise that number.

Morgan Stanley is among them. In a research note earlier this month, the firm's analysts wrote that the AI "ecosystem remains compute-constrained and urgency to spend remains high." They now project Amazon's 2026 capex will actually reach $218 billion, with spending continuing to climb through 2028, when they expect the annual figure to hit $318 billion.

The case for AWS

The bet behind that spending is AWS, and Morgan Stanley made the bull case explicit, pointing to Amazon's "private lab deals" as evidence of what it called multi-year growth durability. Amazon has been steadily deepening ties with the two AI labs everyone in the industry is watching, OpenAI and Anthropic, while also expanding its chip partnerships. In April, it struck a deal to supply Meta with its custom Graviton chips.

Morgan Stanley's reasoning: AWS has access to nearly every major AI model, from large frontier systems down to small, customized ones, which positions it well in a market increasingly focused on squeezing down the cost of running each AI task. As the analysts put it, that access "position[s] it as a winner in a world where optimizing token cost per task is the key."

That thesis will be tested against some genuinely strong competition. Google Cloud revenue jumped 82% year over year in its most recent quarter, and Microsoft's Azure grew 43% in its fiscal fourth quarter. AWS revenue is expected to rise about 31% from a year ago, according to StreetAccount, an acceleration from the 28% growth it posted in the first quarter, which was already its fastest pace in more than three years. Strong, but still a fraction of what Google Cloud just put up.

Cutting jobs while spending big

Even as Amazon pours money into infrastructure, it's been trimming its workforce elsewhere. The company has held layoffs in customer service and seller support in recent months, and last week it announced job cuts within its artificial general intelligence unit, the latest shakeup for a group that recently got a new leader after the head of its AGI Lab announced his departure.

It's a pattern showing up across the industry: aggressive investment in AI infrastructure and models, paired with headcount reductions in more traditional parts of the business.

Prime Day's calendar shuffle

Amazon's retail business also has a wrinkle to explain this quarter. The company moved Prime Day up from its usual July slot, citing a crowded summer calendar that includes the World Cup and the 250th anniversary of American independence. Mizuho analysts flagged in a Monday note that the shift could dent North America retail sales growth in the third quarter before things pick back up in the fourth.

Prime Day ran June 23 through June 26 this year, and Adobe data shows U.S. online spending across all retailers during the event grew about 9.3% year over year to $26.4 billion. Amazon itself doesn't disclose sales figures from the promotion. Evercore analysts called the event "reasonably successful," while KeyBanc's proprietary data showed Prime Week spending up 41.7%, a slowdown from 50.5% growth the year before. KeyBanc's read: strong, just not as strong as last year.

What's riding on Thursday

Amazon shares are up about 3% so far this year, trailing the S&P 500's roughly 8% gain, a sign that investors have been more cautious on the stock than the broader market. Thursday's numbers, and especially any update to the capex outlook, will go a long way toward determining whether that caution eases or deepens. Given how differently the market treated Microsoft and Meta this week for the same basic bet on AI infrastructure, the message from Amazon's cloud growth may matter just as much as the size of the check it's writing to fund it.