Microsoft Cuts AI Data Center Expansion – Here’s Why

Microsoft is scaling back its data center expansion in the U.S., canceling leases for several hundred megawatts of capacity, reports TD Cowen, an American investment bank and financial services firm.

The move suggests the company may be reassessing its long-term AI infrastructure needs.

Canceled Data Center Leases

Microsoft, a leading tech company investing in AI and AI data centers, has canceled select U.S. data center leases and scaled back international expansion, according to a February 21 note from TD Cowen circulating on X.

“Microsoft has canceled select leases with at least two private data center operators across multiple U.S. markets, totaling ‘a couple of hundred megawatts,’” the document states.

The company has also halted finalizing certain data center leases that were still in the early negotiation stages.

Additionally, Microsoft has redirected a significant portion of its planned international investments to the U.S., signaling a slowdown in overseas expansion.

The exact reason behind these moves remains unclear, as Microsoft has yet to provide a statement. However, TD Cowen suggests that the company may be facing an oversupply of data center capacity.

AI Demand Uncertainty

AI models require massive computing power, driving a surge in demand for GPUs, cloud resources, and data centers. Microsoft is one of the biggest spenders on AI infrastructure and the leading company in leasing data center space throughout 2023 and the first half of 2024.

The company plans to commit $80 billion this year to AI data centers. However, recent lease cancellations raise questions about whether it may overestimate future demand.

According to TD Cowen, this shift in capacity needs may be linked to OpenAI. Microsoft, a strategic partner of OpenAI, has paused construction on parts of its multibillion-dollar data center in Mount Pleasant, Wisconsin, which was reportedly intended to support OpenAI.

In January 2025, OpenAI announced Stargate, a significant collaboration with SoftBank and Oracle, to boost computing power and reduce reliance on Microsoft’s Azure. By 2030, OpenAI expects Stargate to provide 75% of its computing power.

Microsoft’s decision to halt construction signals it may have overestimated its data center needs, particularly in regions where excess capacity cannot be easily repurposed for cloud services, raising the risk of oversupply under its revised projections, TD Cowen says.

Tech Giants Pour Billions into AI Data Center Expansion

The AI boom is fueling record-breaking investments in data center infrastructure as major tech companies race to meet the growing demand for AI-driven computing power.

Amazon Web Services is leading the charge, committing more than $100 billion to cloud and AI infrastructure over the next decade. Google is investing $75 billion in AI and cloud computing, rapidly expanding its data center capacity. 

Meanwhile, China’s Alibaba, which recently integrated DeepSeek’s AI models, plans to spend $50 billion over the next three years to scale up its data centers and enhance its AI infrastructure.

The surge in spending underscores the escalating competition in AI and cloud computing as companies seek to establish long-term dominance.

On The Flipside

• Microsoft’s pullback comes just as its competitor, billionaire Elon Musk, who is closely connected to Donald Trump’s inner circle, introduced the Grok-3 chatbot this month through his AI startup, xAI. While specific investment figures for Grok-3 have not been disclosed, xAI has made substantial infrastructure commitments and is reportedly seeking to raise over $10 billion in new funding.

Why This Matters

Microsoft has been a top investor in AI infrastructure and a major data center customer. Its pullback and OpenAI’s shift to other cloud providers signal growing competition, potentially reshaping the AI and cloud market.

Check out trending DailyCoin’s articles:

Bybit Restores Reserves After Massive Hack

Pi Network Launch Sends Pi Coin Price on a Rollercoaster Ride