# Microsoft Targets Massive Expansion of Global Data Centre Infrastructure to Meet AI Demand
Microsoft is embarking on one of the most ambitious data centre expansion plans in the history of the technology industry, with reports indicating the company aims to grow its global data centre capacity to roughly 38 gigawatts by 2032 — a more than threefold increase from its current footprint of approximately 12 gigawatts.
## The Scale of the Expansion
The current data centre capacity of around 12 gigawatts includes a portion dedicated to artificial intelligence workloads. Of that total, roughly 2 gigawatts is powered by AI-specific silicon. Under the expansion plan, AI-centric capacity would grow to approximately one-third of the total 38 gigawatts, meaning around 12.7 gigawatts would be devoted to AI processing — more than six times the current AI allocation.
The overall increase of approximately 26 gigawatts represents a significant leap in global data centre infrastructure and underscores the extraordinary demand being placed on cloud computing by AI workloads.
## Recent Progress and Momentum
During its fiscal fourth quarter, which ended in June, Microsoft added 1 gigawatt of new data centre capacity and deployed 31 new facilities across five continents. Across the full fiscal year 2026, the company added 88 data centres in total.
The company also made notable strides in hardware deployment speed, cutting the time between receiving new GPUs and putting them into service by nearly half during the financial year. This acceleration in capacity rollout has been critical as customer demand continues to outpace available infrastructure.
Azure, Microsoft’s flagship cloud platform, generated over $100 billion in annual revenue during fiscal 2026, representing 41% growth. Cloud services revenue, which encompasses Azure and other commercial cloud products, surpassed $214 billion for the year. In the fiscal fourth quarter alone, Azure revenue grew 43% year-on-year.
## Cutting-Edge Data Centre Design
Microsoft’s Fairwater AI data centre architecture represents a significant engineering achievement. Each rack operates at approximately 140 kilowatts, and each row delivers about 1.36 megawatts of power. The facilities rely on closed-loop liquid cooling systems that reuse coolant after an initial fill, minimising water consumption during normal operations.
The backend networking uses a two-tier Ethernet architecture capable of 800 gigabits per second GPU-to-GPU communication. Microsoft has also constructed a dedicated AI wide-area network to link its AI data centres across geographically dispersed locations.
In parallel, Microsoft is developing its own custom AI silicon alongside processors from partners including Nvidia and AMD. The Maia 200 inference platform, for example, uses a two-tier Ethernet architecture that can scale clusters to as many as 6,144 accelerators and incorporates the company’s second-generation closed-loop liquid cooling hardware.
## Power Infrastructure and Energy Strategy
Meeting the energy demands of this expansion requires a parallel investment in power infrastructure. Microsoft has contracted 40 gigawatts of new renewable energy capacity across 26 countries through more than 400 agreements with utilities and developers since 2020. As of early 2025, 19 gigawatts of that contracted capacity was already operational, with the remainder scheduled to come online over the following five years.
The company achieved its goal of matching annual global electricity consumption with renewable energy in 2025. For new sites like the upcoming Pecos campus in Texas, Microsoft plans to deploy co-located natural gas power facilities that operate behind the meter, delivering electricity directly to the data centre before eventually connecting to the regional grid.
## Broader Implications for the Energy Sector
The scale of data centre growth has significant ramifications for global energy systems. According to the International Energy Agency, data centres consumed approximately 415 terawatt-hours of electricity worldwide in 2024, with projections reaching around 945 terawatt-hours by 2030 under its base scenario.
AI-driven server demand is expected to grow at roughly 30% annually through 2030, accounting for nearly half of the net increase in global data centre electricity consumption. Conventional server electricity use, by contrast, is projected to grow at about 9% annually.
One critical challenge highlighted by energy analysts is the difference in development timelines between data centres and the power infrastructure that supports them. While a new data centre can typically become operational within two to three years, building new power generation facilities and grid infrastructure requires considerably longer planning and construction periods — a mismatch that could create bottlenecks in the years ahead.
## Financial Commitment
Infrastructure investment remains a central pillar of Microsoft’s capital allocation strategy. Capital expenditure reached $41 billion in the fiscal fourth quarter, with approximately two-thirds directed toward shorter-lived assets such as central and graphics processing units. The company also recorded $5.6 billion in finance leases during the quarter, primarily for large-scale data centre sites.
Looking ahead, Microsoft expects capital expenditure of approximately $175 billion for calendar year 2026. An upcoming accounting change will reclassify more future data centre leases from finance leases to operating leases, but the company has emphasised that this adjustment affects only the accounting treatment and not the underlying investment plans.
—
## Frequently Asked Questions
**Q: Why is Microsoft expanding its data centre capacity so aggressively?**
A: The expansion is driven by surging demand for cloud computing and AI services. Customer demand for Azure and other Microsoft cloud products has consistently exceeded available capacity, and AI workloads require enormous amounts of specialised computing power that translates directly into data centre infrastructure needs.
**Q: What makes AI data centres different from conventional data centres?**
A: AI data centres require significantly higher power densities per rack, advanced cooling solutions such as closed-loop liquid cooling, high-bandwidth GPU-to-GPU networking, and specialised silicon. These facilities are purpose-built to train and run large AI models efficiently at scale.
**Q: How does Microsoft plan to power all of these new data centres?**
A: Microsoft is pursuing a multi-pronged energy strategy that includes contracting large volumes of renewable energy, deploying co-located natural gas facilities for new sites, exploring nuclear power through long-term purchase agreements, and investing in grid infrastructure.
**Q: What is the Pecos, Texas data centre campus?**
A: The Pecos campus is one of the largest single capacity additions in Microsoft’s history. It represents a multibillion-dollar development expected to take five to seven years to complete and will add approximately 2 gigawatts of capacity, along with dedicated power infrastructure.
**Q: How does Microsoft’s AI chip strategy fit into this expansion?**
A: Microsoft is diversifying its silicon portfolio by developing its own AI processors alongside accelerators from partners like Nvidia and AMD. Custom silicon allows the company to optimise its infrastructure for specific AI workloads and reduce dependency on third-party supply chains.
**Q: Will all of this expansion actually happen?**
A: The 38 gigawatt target comes from third-party reporting and has not been officially confirmed by Microsoft. However, the company’s demonstrated pace of capacity additions and its public statements about demand exceeding supply strongly suggest significant expansion is already underway and will continue.
—
## Conclusion
Microsoft’s planned expansion to 38 gigawatts of data centre capacity by 2032 represents a transformative moment in the technology industry. The scale of investment — both in physical infrastructure and in energy — reflects the unprecedented demand driven by artificial intelligence and cloud computing. From Fairwater’s liquid-cooled AI facilities to the Pecos campus and a global renewable energy portfolio spanning 26 countries, Microsoft is building the foundation for what could become the largest computing infrastructure programme in corporate history. The challenge ahead lies not only in constructing the data centres themselves but in ensuring the power grids, energy generation, and supply chains can keep pace with this extraordinary growth trajectory.
Thank you for reading
