Microsoft Accelerates AI Infrastructure Buildout, Deploys 1GW of Data Center Capacity in Single Quarter

Microsoft Accelerates AI Infrastructure Buildout, Deploys 1GW of Data Center Capacity in Single Quarter

January 29, 2026

In a significant acceleration of its global infrastructure expansion, Microsoft Corporation brought one gigawatt (GW) of new data center capacity online in the second quarter of its 2026 fiscal year. This massive deployment underscores the intense capital demands of the artificial intelligence era, as cloud hyperscalers race to build the computing power required to train and run advanced AI models.

The tech giant reported that this single-quarter addition of 1GW marks a rapid increase in its deployment pace. For context, the company deployed a total of 2GW throughout the entirety of the previous fiscal year. Alongside bringing this capacity online, Microsoft also committed to new data center investments across seven countries during the quarter, signaling a continued aggressive global buildout strategy.

This infrastructure surge was accompanied by a record quarterly capital expenditure of $37.5 billion. Of this total, approximately two-thirds was allocated to "short-lived assets" like GPUs and CPUs, while data center leases accounted for $6.7 billion. The capex figure represents a sequential increase from $34.9 billion in Q1 2026 and $24.2 billion in Q4 2025.

During the company's earnings call, CFO Amy Hood addressed investor concerns about the soaring expenditures. She noted that capex is expected to "decrease on a sequential basis due to a normal variability from cloud infrastructure build-outs and the timing of delivery of finance leases," but indicated that the proportion of spending on semiconductors would likely remain high. Hood emphasized that Microsoft is making "long-term decisions" for its cloud platform, investing across the entire technology stack to benefit customers.

The quarter also saw key technological developments, including the deployment of Microsoft's custom-designed Maia 200 AI accelerator chips. CEO Satya Nadella stated that optimizing for performance and cost efficiency—measured in "tokens per watt per dollar"—is a priority. He reported that system and software optimizations have already increased throughput by 50 percent for certain AI inferencing workloads. Nadella affirmed Microsoft's ongoing partnerships with chipmakers like Nvidia and AMD, stating the goal is to maintain a fleet with the best total cost of ownership across generations.

Financially, Microsoft posted total revenue of $81.3 billion, a 17 percent year-over-year increase. Its cloud segment revenue crossed a milestone, reaching $51.5 billion (up 26 percent YoY), while Azure revenue grew 39 percent. However, the company reported a substantial remaining performance obligation (RPO) of $625 billion, with an average contract duration of 2.5 years. Notably, 45 percent of this RPO is attributable to its strategic partnership with OpenAI.

The scale of Microsoft's investment highlights a broader industry trend where revenue growth, while strong, is being closely scrutinized against the backdrop of unprecedented capital outlays for AI infrastructure. The company's accelerated buildout is set to significantly expand its global cloud and AI service capabilities in the coming years.

Source: datacenterdynamics