Microsoft announces massive expansion of data center capacity specifically to support AI workloads and LLM inference.

Microsoft is set to triple its datacenter capacity within the next year as a response to the surging demand for artificial intelligence. The surge in demand for foundation models and generative AI has been exponentially adding to the amount of computing power needed. Microsoft is now planning to double its Azure capacity in the second half of this financial year—which equates to the first six months of this calendar year—and triple it in the first six months of the next financial year.

The company's expansion plans come from internal documents seen by Insider, which detail Microsoft's efforts to secure "a record number of GPUs" to bolster its infrastructure for handling AI workloads. Redmond has already doubled its installed base of accelerators in the past year in light of heightened AI demand. More broadly, the company last year committed to spending billions to address data center and cloud infrastructure. Though the company has not explicitly confirmed concrete plans, observations align with the reported intent to grow its data center footprint, with a real estate job ad posted earlier this year alluding to the company's plans to expand its footprint in the Asia-Pacific region.

Current data center distribution shows concentration in specific regions. According to Synergy research, over half of all data center capacity—51%—is located in the US, while another quarter (26%) is located in the Asia-Pacific region, though China accounts for nearly two-thirds (62%) of that quota. Europe lags behind, accounting for 17% of all capacity. Looking ahead, market analytics firms expect data center capacity to double in four years, with generative AI emerging as the driving factor.