Data Center

Generative AI could drive rural data center and broadband growth, according to a new report from CoBank Knowledge Exchange.

Data centers traditionally have been built in urban areas, where there is plenty of power and broadband connectivity. But according to the report, “the processing dynamics of generative AI and the enormous power requirements . . . are turning that model on its head.”

“Data center operators recognize that moving gigabits of data is easier than moving electrons and are locating where land is available and the path to meet their power needs is clear.”

Author and CoBank lead economist Jeff Johnston notes that data center operators are targeting areas of Ohio, Mississippi, Utah and other states.

As Johnston explains, generative AI is driving data center demand to increase dramatically at the same time that the chips underlying AI are requiring considerably more power than the chips traditionally used in data center equipment.

All this drove a 72% increase in data center rental rates in 2023, which in turn, is driving data center operators to move into secondary markets and explore rural areas, according to the report.

Source: JLL Research

AI and Rural Data Centers

As Johnston notes, generative AI has two core elements – training and inference.

“In training, a model analyzes vast amounts of data to develop a knowledge base,” the report notes.

“The model then transitions to the inference phase where it uses the knowledge base developed in training and predicts outcomes based on new data.”

Data centers in secondary and rural markets may be good options for AI training because training isn’t latency dependent. In contrast, inference is more likely to occur close to where more people are using applications. The reason is that the applications typically have stricter latency requirements but are not so compute intensive.

Johnston sees an opportunity for broadband providers to partner with and provide high-speed connectivity to the new data centers in secondary and rural markets.

There may be a secondary benefit for the providers and their communities as well.

“Perhaps this investment could be parlayed into offering fiber-to-the-home in underserved/unserved neighboring towns,” the report speculates.

Join the Conversation

One thought on “Report: AI Could Drive Rural Data Center and Broadband Growth

  1. Another generational solution. The 1980s and 1990s brought huge investments in rural areas first with FIS/American Express, then Acer, then Discover, then Google . All made major installations in rural areas with a perfect match of investment in infrastructure and jobs for the matched skillsets in the region. Just yesterday I was wondering aloud if these sites will be upgraded or duplicated.

Leave a Reply

Your email address will not be published. Required fields are marked *

Don’t Miss Any of Our Content

What’s happening with broadband and why is it important? Find out by subscribing to Telecompetitor’s newsletter today.

You have Successfully Subscribed!