Some analysts expect AI spending to remain a priority even if overall IT budgets get stretched, but others are more skeptical.

One thing is clear after the latest earnings season: Big Tech plans to plow ahead and deploy massive budgets for artificial-intelligence data-center buildouts.

That's even as markets have sold off this year on recession fears, adding to the concern from some on Wall Street that companies could be allocating too much to AI given the technology's scant returns on investment so far. The fear is that companies may have to cut back if the economy goes into a recession. That would presumably affect chip makers like NVIDIA Corp. and other hardware companies that have benefitted from the spending boom.

So far, though, Big Tech isn't backing down. Capital-expenditure growth spending from U.S. hyperscalers accelerated to 71% in the first quarter, with about $81 billion in spending, according to Evercore ISI analysts.

And Alphabet and Microsoft reiterated their capital-spending forecasts this earnings season, while Meta Platforms, Inc. boosted its capital-expense outlook. Amazon.com did not give a capex forecast this week, but confirmed it spent $24.3 billion in the first quarter, mostly to support demand for AI services. In February it projected spending at a pace of about $100 billion this year.

All told, Meta, Amazon, Alphabet and Microsoft still appear to be on track to spend a combined $300 billion-plus this year in building out AI data centers and their infrastructure.

Wall Street seemed content enough with the spending plans, judging by the stock reactions.

Some analysts expect AI to remain a priority even if overall budgets have to tighten.

"This is just such an extraordinary situation," Crawford Del Prete, president of market-research firm IDC, said last month following the fluctuating news from the Trump administration about tariffs. He said he had just been speaking with an information-technology executive who was putting the brakes on spending. "They are stalling, they are freezing," he said.

Just several weeks ago, IDC's forecast for overall IT spending was for it to grow about 9% this year, but now analysts there expect 4% to 5% growth. "There's a general hesitancy associated with the uncertainty that we are experiencing," as a result of the complex tariff situation, Del Prete said. He expects companies will continue to invest in AI infrastructure, but at the expense of other information technology spending, which is expected to slow.

"The AI efforts, they are going to prioritize those," he added noting that companies are instead pausing new initiatives, such as fresh projects or things that change the way they do business. "That is where I have seen the biggest stall, but they are not stalling altogether, they are just pausing." He add that if companies have plans around AI development, "stopping that has a big downstream effect."

Raymond James analyst Srini Pajjuri took a similar view in a note to clients Friday.

"While additional tariffs could impact hardware demand, we expect AI-related spending to be more resilient, given the long lead times and intense competition among hyperscalers," he wrote.

Leading up to earnings this week, there had been plenty of investor concerns and questions in recent notes over the last few weeks. Reports from Bloomberg that both Amazon and Microsoft were pulling back on some early-stage data center projects garnered investor attention, but both companies maintained in LinkedIn posts their unwavering commitment to their spending plans.

"The reality is we've always been making adjustments to build, lease, what pace we build all through the last, whatever, 10, 15 years," Microsoft Chief Executive Satya Nadella added on his company's earnings call Wednesday. He's satisfied with the pace of Microsoft's efforts and says the company takes into account things like power needs, geographical balance and future workload requirements.

Microsoft still doesn't expect to have enough capacity to meet AI cloud demand for the current quarter.

"We think the uptick in capex forecasts should quell any concerns surrounding recent chatter on pauses/moderation in data center capacity demand," Evercore ISI analyst Amit Daryanani wrote.

But even after the earnings reports this week, there are still some concerns that the companies may be overspending, especially given the AI developments in China around products like DeepSeek that are seeking to lower the cost of computing power needed to power AI.

"Google and Meta may decide to remain more aggressive as they increasingly fall behind in AI, but that could very well backfire with investors," said Gil Luria, a D.A. Davidson analyst. "If they stick with their out-of-control capex plans in a potentially weakening economy investors may lose patience."

Meanwhile, Jay Goldberg of Seaport Research Partners launched coverage of Nvidia's stock this week with Wall Street's only sell rating. Among his concerns? "Growing scrutiny of AI budgets at enterprises and direct customers."

Goldberg noted "enterprises searching for compelling AI applications beyond minor 10% to 20% cost reductions in limited areas," adding that "this is a natural function of technology adoption curves, but that curve looks fairly steep right now."