(March 12): Microsoft Corp and Meta Platforms Inc each committed nearly US$50 billion ($63.8 billion) in additional data centre leases in their most recent quarters, underscoring an escalating bet the tech industry is making on artificial intelligence.
Those pledges helped push overall commitments to future data centre leases above US$700 billion among the largest cloud computing companies, a group that also includes Oracle Corp and Amazon.com Inc. These obligations have steadily climbed over the last year as tech giants ink deals to rent server farms, according to a Bloomberg analysis of quarterly filings.
The future costs, which come on top of active leases, won’t appear on companies’ balance sheets until they begin making payments on them. They’re generally tied to data centres but can also include facilities such as offices or warehouses. Some leases include clauses that allow companies to escape future obligations under certain conditions.
Microsoft and Meta have each been vocal in recent months about getting more computing capacity to power the development of AI products. Microsoft carries about US$155 billion in future lease commitments, while the latter company has US$104 billion.
The spending increase is particularly notable for Microsoft, which had taken a pause on its data centre leasing activity through much of 2025. Now, a lack of server farm space has become a major concern for company executives and investors. In the quarter ending in December, Microsoft added over one gigawatt in data centre capacity. That’s roughly equivalent to the power generated by a nuclear reactor.
See also: India plans new US$11 bil fund to support local chipmaking — Bloomberg
Representatives for Microsoft and Meta declined to comment.
Overall, Oracle has committed the most money to data centre rentals. It’s on the hook for US$261 billion in leases that have yet to begun, the company said in a filing Wednesday. Still, its new commitments in the most recent quarter were smaller than some peers, as it had already signed many of the large sites needed for a massive contract with OpenAI.
Uploaded by Magessan Varatharaja
