Microsoft, Google, Meta To Aggressively Raise Data Center Spending — With Mixed Reviews

The AI arms race is showing no signs of slowing down.

Even after more than a year of unprecedented spending on the data centers and information technology equipment to support artificial intelligence, tech giants Microsoft, Google and Meta all indicated this week that they will spend billions more than anticipated on developing AI infrastructure. 

But the three firms’ plans to continue growing development capital expenditures have met differing receptions on Wall Street following their first-quarter earnings reports this week, with investors rewarding Microsoft and Google for evidence that their spending is driving tangible gains for their cloud businesses.

Meta, on the other hand, saw its share price plummet after the firm acknowledged its long-term bet on ramping up AI data center spending is unlikely to yield many short-term benefits — even as CEO Mark Zuckerberg doubled down on the firm’s ballooning capex. 

“We should invest significantly more over the coming years to build even more advanced models and the largest-scale AI services in the world,” Zuckerberg told analysts Wednesday.

“As we're scaling capex and energy expenses for AI, we'll continue focusing on operating the rest of our company efficiently, but realistically ... we'll still grow our investment envelope meaningfully before we make much revenue from some of these new products,” he said.

Microsoft’s capital expenditures reached $14B last quarter, a nearly 22% increase from the previous three months. The firm said it expects infrastructure spending to increase materially throughout the year, putting the company on track to ramp up capex by more than 50% to north of $50B. 

Microsoft’s quarterly guidance tracks with a report published last week by Business Insider that the company plans to double the amount of data center capacity it builds in the first half of this year.

Citing leaked documents, the report suggests Microsoft intends to add a gigawatt of capacity over the next six months, with 1.5 GW planned for 2025. By comparison, Microsoft has around 5 GW of data center capacity today.

Meta is also ramping up its already-ambitious development spending. The company upped its expected annual capex this week to as much as $40B, which would be a significant jump from $28.1B last year. While Meta didn't provide guidance beyond 2024, the firm indicated capex should continue to rise next year “as we invest aggressively to support our ambitious AI research and product development efforts.”

Google parent company Alphabet reported capex increased 9% to $12B last quarter, driven overwhelmingly by servers and data centers. CEO Sundar Pichai told analysts spending on data centers and other digital infrastructure will remain at or above current levels for the rest of the year. 

“We have committed to making the investments required to keep us at the leading edge in technical infrastructure,” Pichai said Thursday. “You can see that from the increases in our capital expenditures. This will fuel growth in the cloud and help us push the frontiers of AI models and enable innovation.”

This spending surge from three of the four largest hyperscale data center users — Amazon reports its quarterly results next week — adds fuel to a data center development landscape already experiencing unprecedented growth. Global hyperscale data center capacity has doubled over the past four years, according to data analytics firm Synergy Research.

As many as 130 hyperscale facilities are expected to come online each year for the next decade, with 440 already in the pipeline — and those data centers are larger than ever before.  

“While the number of operational hyperscale data centers has now passed the 1,000 mark, the bigger news is that overall capacity is growing even more rapidly than the count of data centers,” Synergy Chief Analyst and Research Director John Dinsdale said. “Generative AI technology and services will push this trend even harder over the next five years.”

Wall Street has started to eye this kind of ballooning AI infrastructure spending warily. Investors increasingly want to see AI infrastructure spending driving short-term revenue, with thinning patience for multibillion-dollar long-term bets on AI that may not yield a significant return for years. 

Yet despite rapidly escalating capex, the markets reacted enthusiastically to Alphabet and Microsoft’s quarterly results. Alphabet’s share price surged in the hours after its earnings were released late Thursday, pushing the firm’s valuation past $2T.

In addition to 16% top-line revenue growth that exceeded expectations and the announcement of the firm’s first dividend, investors were buoyed by 30% revenue growth from Google’s cloud business, a significant increase the firm’s leadership attributed to AI and held up as evidence that its bet on AI is bearing fruit. 

But it is Microsoft that has made the clearest case that its massive investment in AI infrastructure is translating directly into short-term value for shareholders. The company’s stock jumped 6% following earnings that showed surging demand for its AI-integrated Azure cloud products driving better-than-expected quarterly revenues of $61.9B. 

Of the major cloud providers, Microsoft was arguably the most aggressive in its pivot toward AI, partnering with OpenAI in 2023 and quickly incorporating the company models into its Azure products.

The result has increased market share, which grew further this quarter, with Azure revenues up 31%. New Azure cloud contracts were also ahead of expectations last quarter, with increases in both deal size and length. Microsoft leadership attributed seven points of growth directly to AI demand, touting it as evidence their high-cost pivot is paying off. 

Indeed, Microsoft’s leadership said its growing data center investment — rather than being speculative — is actually falling behind AI demand. Chief Financial Officer Amy Hood indicated on a call with analysts Thursday that the company can’t bring new data center capacity online fast enough to keep up with demand for its AI cloud products and that these infrastructure limitations may constrain Azure’s growth in the coming month.  

“Currently, near-term AI demand is a bit higher than our available capacity,” Hood said. “That is partially why you see the capital investment in the shape that it is, because right this minute, we do have demand that exceeds our supply by a bit.”

As Wall Street rewarded cloud providers for demonstrating short-term payoffs for their escalating capex, Meta’s stock price tumbled following quarterly earnings that showed little yield from the company’s escalating AI spending. Despite announcing strong quarterly results, with top-line revenue up 27% to $36.5B, shares plunged as much as 19% the following day, knocking as much as $200B off Meta’s market value. 

Of course, the social media giant’s business model and its near-term AI ambitions are much different from those of companies like Microsoft and Google. Meta doesn’t offer its computing infrastructure as a service, instead using its computing power to build its own AI products and services like the newly released Meta AI.

This is a longer-term bet, and Zuckerberg acknowledged that it could be “several years” before Meta sees any significant return on its AI spending. 

It’s a reality that Zuckerberg seemed to anticipate would make investors anxious.

“We've historically seen a lot of volatility in our stock during this phase of our product playbook where we're investing in scaling a new product but aren't yet monetizing it,” he told analysts Wednesday.

“The main thing that I'm focused on for this year and probably a lot of next year is growing AI products and the engagement around them, and I think we should all have quite a bit of confidence that if those are on a good track to scale, then they're going to end up being very large businesses.”