Amazon.com Inc. won over Wall Street this week with improved cloud trends and a more upbeat tone. But perhaps the biggest winner from earnings season was a company that won't deliver its report for another few weeks.

That's Nvidia Corp. (NVDA), which is primed to benefit from the heightened artificial-intelligence spending that management teams at Amazon (AMZN), Alphabet Inc. $(GOOG)$ $(GOOGL)$, Meta Platforms Inc. (META) and Microsoft Corp. $(MSFT)$ talked up this past week.

"The message that I took away is that companies are going to continue to spend on compute," and Nvidia "is probably the best shop right now in terms of supplying it to the marketplace," Matt Stucky, chief portfolio manager for equities at Northwestern Mutual Wealth Management, told MarketWatch.

Mizuho trading-desk analyst Jordan Klein took a similar view, writing in a note to clients that "all roads lead back" to Nvidia, as well as smaller chip players like Broadcom Inc. $(AVGO)$ and Micron Technology Inc. $(MU)$

Meta now anticipates outlaying $70 billion to $72 billion on capital expenditures this year, up from $39 billion last year, and expects its capex growth in dollars to be even higher next year. Amazon now plans to spend $125 billion. And Alphabet is targeting $92 billion this year, up from a prior forecast of $85 billion.

Microsoft doesn't report on a traditional calendar-year cycle but has flipped its messaging, as it now models capex rising at a faster rate in the current fiscal year than last fiscal year's 58% clip. It previously expected the growth rate to slow. The other Big Tech spenders plan to ramp up their expenses next year as well.

"Where those dollars flow, they flow to the compute suppliers, the semiconductor companies, the platforms," Stucky said - and Nvidia is the major one, he added.

Stucky throws that together with commentary from Nvidia's Washington, D.C., iteration of its GTC event on Tuesday - when, in his view, Chief Executive Jensen Huang "basically told you ... that estimates for his business over the next year and a half were about $100 billion light."

Analysts noted that the math is messy, but Bernstein's Stacy Rasgon said Huang seemed to imply the company could do "well over" $300 billion in data-center revenue next calendar year, while the consensus view was for less than $260 billion.

The heavy spending has intensified chatter on Wall Street over whether AI is a bubble, and Stucky acknowledged there are both similarities and differences to the dot-com era.

"The question is whether the end user is going to be able to monetize it effectively, and here the end user is mostly OpenAI. And if you're looking at the unit economics of OpenAI ... that's where the question mark is," he said. "Look at 2026: For every dollar that OpenAI is taking in, $2 is going out the door."

That said, a recent Oracle Corp. $(ORCL)$ bond deal to finance OpenAI-related capacity was heavily oversubscribed, hinting at confidence around OpenAI's ability to improve monetization dramatically, Stucky noted. "But if that were not to be the case, all the spending probably is going to have to slow down significantly."

Tech's pecking order

Nvidia looks to still be in a good spot as hyperscaler spending picks up further, but the latest reports have changed Wall Street's perceptions of other Big Tech players.

Namely, Amazon is getting the Alphabet treatment, with investors starting to see the company's AI positioning in a better light. The Google parent's stock has soared off its spring lows, as Wall Street came to see the company as an AI winner. And now Amazon's stock is coming off a nearly 10% gain on Friday, as the cloud-computing and e-commerce company's latest earnings showed an improved AI story.

Amazon got its cloud growth above 20% in the third quarter - well ahead of the 17.5% rate shown in the second quarter - and the acceleration served as "a proof point that they continue to be not only relevant" but also "the leader in the space," Stucky said.

Investors also seem more willing to put up with swelling AI budgets at Amazon and Alphabet than they do at Meta, partly owing to the differing natures of the businesses and how the companies monetize AI.

For the hyperscalers, "there's less risk to that elevated spend because they're quickly going to monetize that spend in the form of computing," Stucky said. Meta, though, is trying to monetize AI through enhancements to advertising and other elements of its platform, as it pursues what CEO Mark Zuckerberg refers to as "superintelligence."

Zuckerberg is "promising the ROI is going to be there, and it has been - but the further they continue to go, the further the risk is that that ROI number might not be there," Stucky said, referring to returns on investment.

In the meantime, "you get margin and EPS compression" on a year-over-year basis next year, and "have to believe a big [reacceleration] and payoff will come more in 2027," Mizuho's Klein added. He still likes the stock's risk-reward balance but acknowledged that "many will look elsewhere where catalysts and near-term earnings upside could be better."