As AI technology advances, it’s putting ever-increasing strain on existing data centers and associated power sources. To keep up, we’ll need to build a lot more infrastructure.

That’s the broad thinking behind an effort announced today, the Global AI Infrastructure Investment Partnership, that seeks to put up to $100 billion in private equity and debt financing into building and expanding data centers.

Participants include Microsoft, Blackrock, Global Infrastructure Partners and United Arab Emirates AI-focused investment fund MGX. The partnership plans to invest primarily in the U.S., although it may pursue some investments in what it called “partner countries.”

The partnership says it will seek to raise and put to work up to $30 billion in private equity financing, with up to $100 billion in total investment including debt. It will also draw support from Nvidia, which plans to offer expertise in the design and integration of AI infrastructure.

The effort comes amid a boom time for data infrastructure assets. Shares of leading data center REITs, such as Digital Realty Trust and Equinix are up sharply over the past two years. And tech giants like Microsoft, Amazon and Meta are building and planning giant facilities across the U.S.

Just earlier this summer, Microsoft said it spent $55.7 billion on capital expenditures for its fiscal year — up 75% from the prior year — primarily to build data centers.

Power-hungry data centers are also putting increasing pressure on existing energy infrastructure. Per Goldman Sachs, AI is poised to drive a 160% increase in data center power demand by 2030.

Partly that is because AI is an energy-intensive offering. Goldman estimates that on average, a ChatGPT query needs nearly 10x as much electricity to process as a Google search. As AI offerings get more advanced, it’s likely power demands will be even more intensive.

Illustration: Dom Guzman