Microsoft Plans to Triple Global Data Center Capacity to 38 Gigawatts by 2032, AI-Specific Infrastructure to Reach 12.7 Gigawatts

Microsoft Corporation plans to more than triple its global data center capacity from approximately 12 gigawatts today to over 38 gigawatts by 2032, according to people familiar with the plans cited by Bloomberg on September 10, 2026. The build-out covers company-owned and leased facilities but excludes computing power rented from neocloud providers such as CoreWeave. Of the current 12 gigawatts, only about 2 gigawatts is dedicated to AI-specific computing, a share expected to reach roughly one-third of the 38-gigawatt total, or approximately 12.7 gigawatts, by 2032, representing a roughly six-fold increase in dedicated AI compute capacity. The expansion responds to severe supply constraints that have already forced Microsoft to turn away some AI and cloud business, with CEO Satya Nadella emphasizing in July that Microsoft is bringing capacity online faster than ever while implementing strategies to maximize efficiency from existing hardware infrastructure. Microsoft's capital expenditure reached 145 billion dollars last fiscal year, with 329 billion dollars in future lease obligations for data centers. The company expects 50 billion dollars in capital expenditure for the first fiscal quarter of 2027 and anticipates spending 175 billion dollars across the 2026 calendar year. To optimize financial reporting, Microsoft is extending data center lease terms from 15 to 25 years, effectively reducing the annual impact of capital expenses in its reporting. The planned expansion would give Microsoft a computing footprint larger than the peak electricity consumption of New York state.