Big Tech Spending Projections Calm Fears Of Data Center Slowdown

The world’s largest tech companies aren’t slowing down their data center development plans.

Microsoft, Amazon, Google and Meta are collectively ramping up investment in data centers and other infrastructure to support artificial intelligence, their executives said on quarterly earnings calls over the last week, quelling widespread fears of an imminent slowdown in hyperscale data center demand. 

These firms, which together account for the lion’s share of data center demand, all said they plan to either accelerate spending on data centers over the coming year or are sticking with previous plans to grow capital expenditures.

Far from being oversupplied with data center capacity, these hyperscalers reported capacity constraints limiting their revenue growth as demand for AI computing outpaces the development of new data centers needed to support it. 

Big Tech’s bullish capex forecasts present a concrete rebuttal to a pervasive narrative of uncertainty around hyperscale data center demand. Amplified by reports that Microsoft and Amazon have canceled leases and paused new data center projects, some skeptics have sounded alarm bells that tech firms are poised to pump the brakes on their previously insatiable appetites for AI infrastructure. 

But if a slowdown is coming, the companies driving the data center building boom insist it isn’t happening anytime soon. 

“Worries around AI demand following the reports of data center lease cancellations were calmed,” Raymond James analyst Andrew Marok wrote in an investor note following Microsoft’s earnings call Wednesday. “AI demand remains robust.”

Among the hyperscalers, Meta appears to be leaning in the hardest on its data center spending spree. The firm told investors last week it is boosting capex by more than 10%, pushing its annual capital spending from an estimated range of between $60B and $65B up to between $64B and $72B. The firm’s leadership attributed this increase to additional data center investment to support AI, as well as an increase in hardware costs. 

Google parent Alphabet similarly indicated it is staying the course on its accelerated data center spend, with no intention of cutting back plans to ramp up capex to $75B next year.

Amazon reported Q1 capex leaping to $24.3B from $13.9B last year, with the majority going toward data centers and related infrastructure. While the firm didn't provide full-year spending projections, its leadership indicated that its pace of spending will continue, as capacity constraints are limiting its Amazon Web Services cloud unit’s ability to capture additional revenue despite unprecedented demand. 

Despite its highly publicized lease cancellations, Microsoft said it plans to increase its capex next quarter. While its planned annual spending remains unchanged at more than $85B, Chief Financial Officer Amy Hood said she expects capex to increase in the next fiscal year. Like AWS, the firm anticipates capacity constraints will limit growth this quarter due to a shortage of data center space. 

“While we continue to bring data center capacity online as planned, demand is growing a bit faster,” Hood said on the companhy's earnings call. “Therefore, we now expect to have some AI capacity constraints beyond June.”

Beyond accelerated spending, these firms' Q1 earnings also showed growth in the overall market for cloud services, an indicator that AI adoption is picking up steam. The cloud market’s underlying growth rate was close to 25%, according to Synergy Research Group, 7% higher than the growth rate in mid-2023. 

Big Tech’s insistence that its data center spending spree will continue apace — as well as the strong cloud revenue numbers — runs counter to a growing narrative of skepticism about whether bullish demand projections for data center capacity will ever come to fruition.

Fears of an oversupplied data center market first gained traction in January when the release of a new AI model from Chinese firm DeepSeek sent Wall Street into a tailspin. The model's low price tag and development on fewer, more primitive chips cast doubt on an assumption underpinning Big Tech’s AI spending: that massive amounts of energy-intensive chips are needed to develop more advanced AI models that would lead to commercially viable products.

DeepSeek’s success seemed to suggest that it will take far fewer chips — and potentially fewer data centers and less power — to achieve these aims. And a series of headlines over the next three months added further fuel to fears of an imminent data center drawdown from the major hyperscalers.

In March, Joe Tsai, the billionaire chairman of Chinese conglomerate Alibaba Group, warned that a bubble is forming in the data center sector. He said the rush to build data center infrastructure is creating a risk of oversupply in the U.S. as the pace of construction exceeds any realistic demand projections.

The next day, TD Cowen reported that Microsoft had canceled leases or paused construction on gigawatts of data center capacity in the U.S. and Europe as it slows parts of its expansion. The pullback came on the heels of earlier lease cancellations by Microsoft, decisions TD Cowen said “point to data center oversupply relative to its current demand forecast.” In April, Wells Fargo reported that Amazon was also “pulling back” from part of its pipeline of colocation data center lease deals.

Combined with recession fears and macroeconomic uncertainty, these developments prompted some investors and Wall Street commentators to raise doubts about whether hyperscalers’ aggressive spending on data centers will continue. Commentators like Wall Street Journal columnist Asa Fitch voiced the belief that Big Tech firms were already “pretty far over their skis” when it comes to data center spending. 

“Sustaining the boom hinges on the tech companies’ willingness to stick with current and future spending plans against a rapidly weakening economic backdrop,” Fitch wrote last month. “That is a lot to ask. Probably too much.”

While top tech leaders like Microsoft CEO Satya Nadella and Amazon CEO Andy Jassy have consistently pushed back on the notion of diminishing data center demand since the DeepSeek news broke, the capex numbers the firms provided with their Q1 earnings now lend their comments more weight. 

On its earnings call Wednesday, Nadella addressed Microsoft’s decision to exit several lease agreements and pause certain data center development projects. The changes weren't a response to an overall slowing of the firm’s need for data centers or computing capacity, he said. Rather, the canceled capacity represents a shift in where the company’s data centers need to be located as the AI landscape matures and evolves. 

Microsoft doesn’t need fewer data centers, he said. It just needs them in different places.

“The key thing for us is to have our builds and lease be positioned for the workload growth of the future,” Nadella said. “There's a demand part to it, there is the shape of the workload part to it, and there is a location part to it. So you don't want to be upside down on having one big data center in one region when you have a global demand footprint.”

While leaders of the four top hyperscalers all made clear that their capex plans aren't written in stone amid tariff uncertainty and macroeconomic volatility, Big Tech’s Q1 earnings gave Wall Street analysts broad reassurance that spending on data centers and other AI infrastructure is likely to continue accelerating at least into 2026.

Skepticism around demand is unlikely to disappear, but analysts like Evercore ISI’s Amit Daryanani said hyperscalers successfully tamped down fears that the data center building boom is about to grind to a halt.

“We think the uptick in capex forecasts should quell any concerns surrounding recent chatter on pauses/moderation in data center capacity demand,” Daryanani wrote, according to Morningstar.