Microsoft has flipped the switch on a new class of AI data center designed to work not as a single facility, but as part of a continent-spanning compute network that behaves like one machine.
In a new update, Microsoft says its Fairwater family of AI data centers aims to be directly interconnected with data flowing between sites extremely quickly.
According to the tech giant, the Atlanta AI data center is the second site in its Fairwater family, following the construction of its Mount Pleasant, Wisconsin facility.
The Fairwater locations share the same chip, rack and cooling architecture and are connected by a dedicated AI Wide Area Network intended to allow data to move between sites with minimal congestion. Microsoft says it has deployed 120,000 miles of dedicated fiber to enable data to move close to the speed of light.
Microsoft CTO Mark Russinovich says the level of compute needed for training current frontier models now spans multiple sites instead of one.
“The amount of infrastructure required now to train these models is not just one datacenter, not two, but multiples of that.”
Scott Guthrie, Microsoft executive vice president of Cloud + AI, says the connected sites are designed to function as a unified system to support training, fine-tuning and other stages of large AI workloads.
“By making our AI sites operate as one, we’re able to help our customers bring breakthrough models to life, deliver results that matter in the real world and empower them to solve challenges and create new opportunities.”
The tech titan says the Fairwater family of AI data centers will feature a unique design, including Nvidia GB200 NVL72 rack-scale systems, a two-story layout for higher GPU density, and liquid cooling that uses almost zero water.
Microsoft says the data centers in Atlanta and Wisconsin will eventually be connected to facilities under construction across the country.
Disclaimer: Opinions expressed at CapitalAI Daily are not investment advice. Investors should do their own due diligence before making any decisions involving securities, cryptocurrencies, or digital assets. Your transfers and trades are at your own risk, and any losses you may incur are your responsibility. CapitalAI Daily does not recommend the buying or selling of any assets, nor is CapitalAI Daily an investment advisor. See our Editorial Standards and Terms of Use.