The company handily beat Wall Street analysts’ consensus forecasts for both revenue and profit growth, and provided more buoyant guidance for its next quarter than expected too.

Microsoft shares jumped on the news, gaining as much as 9% in after-hours trading on Wednesday evening. The boost was a relief to shareholders, who have weathered a 5.5% slide in the company’s shares so far this year.

The news is also likely to be a comfort to the entire AI sector—evidence that companies are not souring on the much-hyped technology.

It remains to be seen whether inertia explains Microsoft’s strong performance.

Cloud computing contracts cannot easily be canceled. The company may yet see a negative impact from the economic uncertainty U.S. President Donald Trump’s tariffs have caused.

But it could well be that even in the face of a tariff-induced downturn, cloud computing and artificial-intelligence spending will prove resilient, with organizations prioritizing spending on digital transformation and automation even as they cut back on things like business travel and advertising.

“When it comes to cloud migrations, we saw accelerating demand with customers in every industry from Abercrombie & Fitch to Coca-Cola and ServiceNow, expanding their footprints on [Microsoft’s cloud computing platform] Azure,” Microsoft CEO Satya Nadella told analysts on the company’s earnings call yesterday.

Meanwhile, Amy Hood, Microsoft’s chief financial officer, said that “demand signals” across the company’s business lines had shown no signs of weakening in the past month, an indication that Microsoft may not take a hit from Trump’s tariffs.

Dan Romanoff, an analyst at Morningstar who covers the company, noted that the company’s commercial bookings grew 17% year over year when measured in constant currency, and that “remaining performance obligations”—the amount of money in cloud computing contracts that the company has booked but that have not yet been recognized as revenue—increased an impressive 34% year over year to $315 billion.

There was also no indication anywhere in Microsoft’s financials of any general disillusionment with AI on the part of the company’s customers, despite news reports that some businesses have found it difficult to achieve the productivity gains and return on investment they hoped the technology would produce.

On the contrary, Microsoft saw revenue at its Azure cloud unit grow 33% in the quarter, well ahead of analyst forecasts of 29%. The company said 16% of that gain could be attributed to AI-related spending, up from 13% in the prior quarter.

The company also reported it had seen a major new cloud commitment from OpenAI. That seemed to belie news reports of significant tension in the partnership between Microsoft and the ChatGPT creator, in which Microsoft is a major investor.

Although Microsoft retains a right of first refusal on any additional data center capacity the AI startup needs, OpenAI has turned to Oracle to provide the cloud computing infrastructure for Project Stargate, OpenAI’s partnership with Japan’s Softbank that will spend $500 billion on new data centers in the U.S. by 2029.

Microsoft’s decision not to provide data centers for Project Stargate, combined with reports that Microsoft had cancelled plans to lease data center capacity with a number of providers, including CoreWeave, had led to speculation that Microsoft was seeing weaker demand for AI-related software and services than it had previously anticipated.

But in yesterday’s earnings call, the company said it has spent $21.4 billion on capital expenditures in the past quarter, a figure that includes leasing data center capacity, and that it was still on track to spend $80 billion on capex for the year. Hood told analysts that capital spending would continue to rise next year too, although the rate of increase would slow from the 57% clip that the company is expected to report for the current fiscal year, which ends in June.

Nadella told analysts that the company is adjusting where it leases data centers to try to meet demand with the right kind of infrastructure in the right geographies, and that this was the reason for some of the cancellation of data center leases, not that overall AI demand was slowing.

He seemed to suggest that a trend towards companies using smaller AI models and AI models whose capabilities are derived from spending more computing power at the time of inference as opposed to during the initial training of the model, was causing the company to rethink what size data centers it needed and where they should be located. Those developments, already underway, had been spotlighted by the debut of the R1 AI model from Chinese startup DeepSeek earlier this year.

“You don’t want to be upside down on having one big data center in one region when you have a global demand footprint,” he said. Hood added that decisions on data centers involved a “very long lead time”—between two and seven years, she said—“so we’re constantly in a balancing position as we watch demand curves.”

Jason Wong, distinguished vice president analyst at tech research firm Gartner, forecast that generative AI spending would remain strong through 2028. He also noted that Microsoft has not had enough data center capacity to fulfill all the AI demand from its customers so far and is still trying to build out enough capacity to meet that demand. “Microsoft Azure has been capacity constrained over recent years and their investments have always been for the long term,” he said.

The buildout in data centers has continued to pressure Microsoft’s profit margins. The company said its gross margin is 69%, down one percentage point from a year earlier. The fall is attributed to the company’s spending on AI infrastructure. But for now, investors seem untroubled by the slight margin compression.

Microsoft’s total revenue for the past quarter topped $70 billion, 2.3% ahead of consensus forecasts. Operating income was $32 billion, 6% ahead of analysts’ consensus projections, while earnings per share were $3.46, well ahead of the $3.22 analysts had predicted.

Update, May 1: This story has been updated to include comment from Gartner analyst Jason Wong.