Microsoft (NASDAQ: MSFT) warned on Wednesday that capacity constraints may impact its AI services as early as June, with growing demand threatening to outstrip current infrastructure — even as the company continues to scale its global data center footprint.
Speaking during the company’s fiscal third-quarter earnings call, CFO Amy Hood said Microsoft is experiencing stronger-than-expected usage of its AI-powered services, and acknowledged the company may face “a little tight” capacity conditions exiting the current fiscal year.
“We had hoped to be in balance by the end of Q4, but we did see some increased demand,” Hood told analysts. “So we are going to be a little short, a little tight as we exit the year.”
AI Growth Outpaces Buildout Timelines
The company’s artificial intelligence platforms — from Azure OpenAI Services to Copilot in Microsoft 365 — have seen rapid adoption in both enterprise and developer ecosystems. But Microsoft’s massive investment in data centers, while ongoing, has not fully caught up to the rising utilization rates.
Microsoft has committed to spending $80 billion on data centers in 2025, with half of that investment dedicated to U.S.-based infrastructure. CEO Satya Nadella noted that the company opened new facilities across 10 countries and four continents this past quarter alone.
Despite this, AI compute demand continues to challenge provisioning timelines. Hood highlighted the multi-year lead times involved in data center development.
“From land to build-out, it can be lead times of five to seven years, two to three years,” Hood said. “So we’re constantly in a balancing position as we watch demand curves.”
Lease Cancellations Raise Questions
The warning comes amid recent reports of Microsoft canceling multiple data center leases. In February, TD Cowen reported that Microsoft pulled out of deals equivalent to “a couple hundred megawatts” — the power footprint of roughly two standard-scale data centers.
While some industry analysts speculated these moves were related to shifting AI infrastructure needs or zoning delays, Microsoft maintains that the lease cancellations are unrelated to the current capacity warning.
“We’re still committed to our original investment targets,” the company emphasized during the call, pushing back on concerns that cost-cutting or overcapacity in some regions could undermine growth plans.
Balancing Immediate Demand with Long-Term Strategy
The tension between near-term AI demand spikes and the long buildout cycle underscores the broader infrastructure challenge facing hyperscale cloud providers. Microsoft, Google, Amazon, and others are racing to provision high-performance compute capacity — particularly for GPU-intensive workloads driven by large language models (LLMs) and other generative AI applications.
Analysts expect Microsoft to continue prioritizing AI-related capital expenditures throughout 2025. The company’s ability to manage capacity bottlenecks — without sacrificing service quality or customer growth — will be a key metric for investors watching the sustainability of its AI-led cloud growth narrative.
Outlook and Investor Sentiment
Despite the warning, Microsoft shares remained stable in after-hours trading, buoyed by strong earnings and robust cloud revenue. Azure, which now represents a significant portion of Microsoft’s total growth engine, continues to outperform most analyst projections.
Microsoft is expected to provide further updates on AI workload prioritization, data center construction progress, and regional service expansion in its upcoming fiscal Q4 guidance, scheduled for release in late July.
