Microsoft Triples Data Center Capacity to 38 GW by 2032 After Turning Away AI Business

Internal plans at the software giant target a three-fold increase in global cloud infrastructure to accommodate growing artificial intelligence demand. Microsoft will grow its data center capacity from roughly 12 GW today to more than 38 GW by 2032, adding 26 GW of new capacity to its fleet. AI specific compute will expand from 2 GW to about 13 GW, or one third of total capacity, a figure that excludes rentals from neocloud providers like CoreWeave.

Capacity shortages recently forced Microsoft to restrict certain cloud subscriptions and turn away AI business. The company spent $145 billion in capital expenditures last fiscal year and holds $329 billion in future data center lease commitments. This infrastructure buildout extends beyond accelerator clusters, as newer AI tools require increased CPU capacity to complement GPUs.