- Microsoft execs said the hyperscaler is still behind on meeting AI demand
- Capacity constraints are expected to continue through the end of this year
- Analysts think all hyperscalers will face a short term AI capacity shortage, which will benefit smaller cloud providers but benefit the big players in the long run
It may have just posted $22 billion in profit, but Microsoft has something of a problem on its hands. For the second quarter in a row, company executives admitted that it's behind on meeting demand for artificial intelligence (AI) capacity. And that’s not exactly about to change.
“We are constrained on AI capacity. And because of that, actually, we ... have signed up with third parties to help us as we are behind with some leases on AI capacity,” CFO Amy Hood said in response to an analyst question during the company’s earnings call. She added Azure and AI capacity constraints are expected to persist at least through the back half of 2024.
Like Google Cloud, Microsoft is throwing money at the problem.
Hood said it shelled out a whopping $19 billion in capital expenditures – which New Street Research noted is 93% higher than the $10.7 billion it spent in its fiscal Q4 2023 – in the recent quarter. Nearly all of that $19 billion went toward cloud and AI, she noted.
But the data centers which will house its cloud and AI compute take time to build. So, what does the company plan to do in the interim?
Well, as Hood mentioned, it’s teaming up with third parties like Oracle to help it meet demand.
Industry-wide issue
Patrick Moorhead, founder and principal analyst at Moor Insights and Strategy, told Fierce that he doesn’t think the capacity problem is a Microsoft-only issue. Instead, Moorhead said he believes there will be an “industrywide computing challenge for AI in the next six months.”
And that’s good news for smaller, competitive cloud players who might have some available compute on their server shelves.
“Whomever has the capacity and power will get the business” at least in the short term, he said.
“OCI has been a huge beneficiary as has CoreWeave,” Moorhead added.
Moorhead noted there’s one lingering question: whether “AWS and Google have an advantage with their own homegrown silicon. Microsoft has Maia but it’s relatively new.”
In any event, New Street Research seems to think the oversubscribed demand is good news for Microsoft’s long term trajectory.
“Azure growth will accelerate in 2HFY25 supported by AI capacity,” Pierre Ferragu and team wrote in a note to investors. “Gen AI enhances Microsoft’s long-term growth outlook. The ramp of capital-intensive AI in the revenue mix lowers returns on tangible assets, putting minimal and transitionary pressure on Free Cash Flow and Ebit margins, but maintaining overall group profitability in the long run.”
Earnings snapshot
Fiscal Q4 2024 revenue of $64.7 billion jumped 15% year on year with profit up 10% to $22 billion. Hood said server products and cloud services revenue grew 21% while Azure and other cloud services revenue increased 29%, but didn’t provide breakout dollar figures for either. AI contributed 8 points to Azure’s growth, she said.