The impacts of generative AI and data centers used to house these servers have a large impact on the Great Lakes region than some people may think. The Alliance for the Great Lakes, a nonpartisan nonprofit working across the region to protect the natural waters of the Lakes, is one of many organizations who are working to help inform the public of such matters.  

This past week, the Alliance held a webinar on Preparing the Great Lakes Region for Large Water Users. This informational session discussed how large water-using sectors, such as data centers, agriculture, and critical minerals mining, are growing rapidly in the Great Lakes region, and how states are not prepared for these new demands on our water. These demands are leading to unsustainable growth and conflicts over water use, which sooner rather than later will affect boaters both at home and on the water.   

Experts from several organizations joined the call, hosted by Alliance Communications Director Tom Fazzini who moderated the session in which around 700 people attended. The first speaker was Helena Volzer, Senior Source Water Policy Manager for the Alliance for the Great Lakes and author of the Alliance report titled "A Finite Resource: Managing the Growing Water Needs of Data Centers, Critical Minerals, Mining, and Agriculture" in which the topic was focused. In her section of the session Volzer the core of the report and its central focus on where the larger water use increase was stemming from: data centers, critical mineral mining and agriculture. Though all factors have a part to play, data centers have been one of the largest increases in recent years due to the increase demand for generative AI.  

All screenshots are from the "Preparing the Great Lakes Region for Large Water Users" presentation.

This image indicates data centers located in the Great Lakes region.

“A data center using 1 million gallons a day of water for a year is the equivalent of 12,000 Americans would use in that time,” Volzer says. And while different cooling methods of these data servers were noted and have been explored, overall the total footprint of how much water is being used is unknown. Less than one-third of data centers are tracking their water use, as according to Volzer about 97% of data center operators are connecting to municipal supply which are not always able to confirm exact sourcing percentages. Other factors including climate were one, as Volzer noted in a recent report that this past August was the driest August since 1895. Higher temperatures in summer months can affect groundwater, lakes, rivers, etc.  

João-Pedro Ferreira, Regional Economist working with Weldon Cooper Center of Public Service and the University of Virginia, followed up on the activity in rising numbers of data centers throughout the country, 20 percent sitting in states around the Great Lakes region. Ferreira notes that this is bound to change, as the growth pattern is bound to increase to 21 new data centers to be built per year in the Great Lakes area over the next 4 years according to contracts. These centers will more likely become hyperscale data centers, which will mean a sharp acceleration in energy demand. Interestingly, though the construction of these data centers is noted to increase, the employment benefits will be lower than originally projected and communities will pay higher prices without or with less benefits to the local economy. In Virginia’s data centers, about 70,000 jobs are created, but 60,000 of those are for construction and therefore prolonged jobs are much lower. “These are sources that don’t demand a lot of local employment,” Ferreira notes.