Microsoft Warns of Potential AI Service Disruptions Due to High Demand
During Microsoft’s recent earnings call, an executive raised concerns about the possibility of disruptions in AI services as the company struggles to keep up with the growing demand without enough data centers available.
Microsoft’s Executive Vice President and CFO, Amy Hood, highlighted the potential challenges ahead during the company’s third-quarter earnings call for fiscal year 2025. She pointed out that there might be constraints on AI capacity as early as June due to increased demand.
“We had anticipated reaching a balance by the end of Q4, but the surge in demand throughout the quarter has put us in a slightly tight spot as we approach the end of the year,” Hood explained.
The timing of Hood’s statement is particularly noteworthy in light of reports that Microsoft has been canceling multiple data center leases this year.
Earlier this year, investment bank TD Cowen revealed that Microsoft had terminated several data center leases equivalent to a significant amount of power, which could have powered two data centers. Subsequent reports further confirmed additional cancellations of data center leases.
However, Microsoft has clarified that these lease cancellations are not necessarily linked. The company reaffirmed its commitment to investing $80 billion in data centers this year, with half of the funds allocated to U.S.-based facilities.
Hood emphasized that current demand does not necessarily reflect future demand, highlighting the long lead times involved in decisions related to data center investments.
Techcrunch Event
Berkeley, CA
|
June 5
BOOK NOW
“These decisions have very long lead times, from acquiring land to completing construction, which can take anywhere from five to seven years,” Hood noted. “We are constantly balancing our investments based on evolving demand trends.”
Microsoft CEO Satya Nadella also highlighted the company’s expansion efforts, noting the opening of data centers in 10 new countries and four new continents during the previous quarter.
