Microsoft signs $130 billion in data centre leases for AI
Microsoft has secured over $130 billion in new data centre leases to expand Azure's AI computing capacity.
AI progress runs on real estate, power and chips. Microsoft just booked a lot more of all three.
Microsoft reported more than $130 billion in new data centre leases in its latest quarter. As of the quarter ended 30 June 2026, total lease commitments that have not yet commenced stood at $329.1 billion, up from $196.6 billion in the prior period.
The company said these commitments are primarily for data centres and, in some cases, depend on contractual conditions. The future costs will not appear on the balance sheet until lease payments start. This helps explain why obligations can jump before they are reflected in reported expenses.
A record jump in AI infrastructure
This was Microsoft’s largest quarterly increase in data centre leasing so far. It follows a slower stretch through much of 2025, when power availability and project readiness tempered activity. The new deals suggest Microsoft is moving quickly to secure space and power for AI training and inference across key Azure regions.
Leasing offers speed and flexibility. It lets hyperscalers add capacity faster than building from scratch. It also allows them to phase commitments to match the timing of AI demand while sharing risk with specialist developers that line up land, permits and grid connections.
How the commitments stack up
The $329.1 billion figure covers leases that have been signed but not yet commenced. It sits alongside active leases already in operation. The jump from $196.6 billion in the previous period shows how much capacity Microsoft secured within a single quarter.
Many projects will come online in stages over several years as power, cooling and fit-out become available. Some agreements include milestone triggers, which is common in large, multi‑year data centre contracts.
What this means for the market
The scale of these leases points to a sustained build cycle for AI infrastructure. Expect continued investment in servers, networking, advanced cooling and on‑site energy solutions, alongside deeper work with utilities and local authorities.
For customers, the goal is clear. Microsoft is aiming to expand Azure capacity so organisations can run heavier AI workloads with more reliability. For the wider ecosystem, the message is equally direct. AI is shifting from pilots to production, and the race to secure power and space is intensifying.