Redmond, US, October 30, 2025- Microsoft said it plans to double its global data-centre footprint within the next two years, underscoring how rapidly growing artificial intelligence workloads are reshaping its infrastructure strategy. The move coincides with what the company described as the “next chapter” in its partnership with OpenAI, aimed at building shared capacity for large-scale model training and cloud-based AI delivery.
In a company blog post published this week, Microsoft said the expansion will focus on regions where power and connectivity can support hyperscale buildouts, adding that the plan reflects “unprecedented demand” for AI-ready cloud infrastructure. The commitment builds on earlier investment disclosures, including more than USD 80 billion earmarked for new data centres and energy projects globally.
Chief executive Satya Nadella told analysts that Microsoft’s infrastructure growth “is now the fastest in the company’s history,” and that capacity will “more than double” by 2027 to meet the needs of Azure and OpenAI customers.
The statement comes as Microsoft continues to deepen integration with OpenAI following its extended partnership announced on October 28. The companies said they will expand joint development of training clusters, inference capacity, and model deployment platforms. Microsoft described the collaboration as central to its plan to “ensure AI is available safely and at scale” through Azure’s distributed network.
Industry analysts say the decision to double data-centre capacity highlights how competition among hyperscalers has shifted from cloud software features to energy-efficient compute. Much of Microsoft’s new capacity is expected to use advanced cooling, modular designs, and on-site renewables. Although the company has not released a regional breakdown, filings in Ireland, the UK, and northern Sweden show ongoing activity consistent with its expansion roadmap.
Nadella said Microsoft’s AI business is now driving “a new era of infrastructure buildout comparable to the early days of the internet,” noting that customer demand for large model hosting, inference APIs, and Copilot services has exceeded internal forecasts. The company’s statement added that every dollar invested in AI infrastructure “returns multiple times in digital productivity” across enterprise and consumer markets.
Microsoft did not specify total site numbers or locations for the new facilities but said updates would be shared as regional permits and power agreements are finalised over the next 24 months.
.jpg)