Microsoft Reportedly Targets 38GW Data Center Capacity by 2032 to Power AI Growth

According to reports, Microsoft plans to more than triple its total data center capacity from 12 gigawatts (GW) to 38GW by 2032. Up to one-third of this target capacity (12 to 13GW) will be dedicated to specialized AI workloads, compared to just 2GW today. Recent compute shortages have bottlenecked Microsoft Azure, preventing it from satisfying surging customer demand and forcing some workload offloading to competitors. Expanding dedicated AI infrastructure directly addresses these capacity constraints and helps Microsoft maintain its leadership position in cloud and AI computing. The 38GW figure includes both Microsoft-owned data centers and capacity leased from traditional partners, but excludes specialized 'neocloud' GPU providers due to construction timelines and uncertainty. The long-term expansion remains subject to change as server cluster builds take years and market preferences or technological shifts could alter requirements.