Microsoft Plans to Triple Data Center Capacity to 38 Gigawatts by 2032

Microsoft plans to expand its global data center capacity from the current 12 gigawatts to approximately 38 gigawatts by 2032. [1] Under this long-term plan, around one-third of the projected 38-gigawatt capacity could be allocated to artificial intelligence infrastructure. [1] Currently, only about two gigawatts of the company's existing capacity is dedicated to AI-specific chips. [1]

To support this growth, Microsoft expects capital expenditure to reach $175 billion for the 2026 calendar year and $50 billion in the first quarter of fiscal 2027. [1] To manage these costs, the company is considering structuring data center leases over 25 years rather than 15-year periods. [1] This strategy would spread expenses over a longer duration and lower the company's annual reported capital expenditure. [1]

This expansion underscores the massive physical infrastructure required to sustain enterprise AI adoption at scale. By moving from a 15-year to a 25-year lease model, Microsoft is attempting to balance the immediate financial burden of its $175 billion annual capital expenditure with the long-term payoff of AI integration. Expanding AI-specific capacity from 2 gigawatts to over 12 gigawatts highlights a fundamental shift in hyperscaler architecture toward specialized hardware. What the sources don't address: How this massive increase in power requirements will be sourced, particularly given the growing constraints on global energy grids.