Microsoft Activates Atlanta Fairwater Data Center, Linking to Wisconsin for AI Superfactory Network

Nov 13, 2025
Microsoft has opened its Atlanta Fairwater AI data center, connecting it to a Wisconsin facility via a dedicated AI network to form part of a distributed supercomputer system.

Microsoft has brought its new Atlanta Fairwater data center online, reports Data Center Dynamics. The two-story facility, located in Georgia, is the second site in Microsoft’s Fairwater network and connects with the first in Wisconsin through a dedicated AI Wide Area Network using fiber-optic links.

The Atlanta site uses a closed-loop liquid cooling system and houses hundreds of thousands of Nvidia GB200 and GB300 GPUs. Each rack supports up to 72 Blackwell GPUs connected via NVLink, with 800 Gbps GPU-to-GPU connectivity across an ethernet-based backend network. Microsoft’s SONiC operating system manages network switches, and a custom Multi-Path Reliable Connected protocol was developed with OpenAI and Nvidia to optimize routing.

The company designed the two-story layout to minimize the distance between racks and opted to omit traditional on-site power generation, UPS systems, and dual-corded distribution, citing Atlanta’s reliable grid. This approach reduced deployment time and costs.

Microsoft plans to connect additional Fairwater facilities to form a distributed supercomputer capable of running large-scale AI workloads across multiple sites simultaneously.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Silicon Brief or Daily AI Brief.

Also, consider following us on social media:

AI Chips & Datacenters AI Brief AI Brief (X)
Free newsletter

Silicon Brief

Weekly coverage of AI hardware developments including chips, GPUs, cloud platforms, and data center technology.

AI Token Webinar

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more

You may also like

Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.