Bloomberg reports that Microsoft plans to expand its owned and leased data-center capacity from around 12 gigawatts today to more than 38 gigawatts by 2032, with Reuters picking up the figures. AI-dedicated capacity would rise from about 2 gigawatts now to roughly a third of that 2032 total. The reporting frames the buildout as a response to a compute shortage that had forced Microsoft to decline some AI and cloud orders. The roadmap covers company-owned and leased sites but excludes rented capacity from smaller cloud providers. Microsoft's capital spending reached $145 billion in its last fiscal year, with substantial future lease obligations attached to data centers.