Microsoft is planning to expand its data center capacity from approximately 12 gigawatts today to more than 38 gigawatts by 2032, a move that would triple the company's computing power, according to Bloomberg1,2.
About a third of the planned 38 GW is expected to be centered on AI-specific chips. The expansion comes as shortages have forced Microsoft to turn away some AI and cloud business.
The jump from 12 GW to 38-plus GW represents roughly 26 GW of new data center capacity over the next six years. ANALYSIS Dedicating approximately one-third of the total footprint to AI-specific chips would concentrate a substantial share of the buildout behind AI workloads.
The disclosure that Microsoft has already had to turn away AI and cloud customers frames the buildout less as speculative positioning and more as a response to demand that has outstripped existing infrastructure.
Microsoft's infrastructure push lands during a period of broader activity for the company on AI governance and legal fronts. On September 9, Microsoft joined the American Federation of Teachers and the United Federation of Teachers in announcing a National AI Safety and Privacy Standard for U.S. school districts[1]. Separately, the Seattle Times and Newsday filed suit against Microsoft and OpenAI on September 5, alleging unauthorized use of journalism to train AI models[2].
Bloomberg first reported the capacity expansion plans.