How Agriculture and Data Centers Compete for the Great Lakes’ Most Precious Resource

Agriculture and data centers are rarely mentioned in the same breath. One is the backbone of the food system and deeply woven into the fabric of the Great Lakes region. The other is a fast-growing infrastructure at the core of our digital world, storing, processing and delivering the data that powers everything from artificial intelligence to social media. But both are competing for the same dwindling, finite resource: freshwater.

The Great Lakes hold a staggering 20 percent of the world’s surface freshwater. Yet despite their size, less than 1 percent of that water is naturally replenished each year. The region’s limited renewal rate makes it uniquely susceptible to overuse, a vulnerability that’s increasingly apparent as demand climbs.

A recent report by the Alliance for the Great Lakes warns that three rapidly expanding sectors — agriculture, data centers and mining — are escalating stress on the region’s water supply. At the same time, climate change is worsening this strain, disrupting rainfall patterns, drying out growing seasons and altering how groundwater is replenished.

Agriculture is pulling more water for irrigation to cope with hotter, drier growing seasons. Much of that water goes to feed crops like corn and alfalfa, which are primarily grown to support livestock production. Data centers are drawing millions of gallons annually to stay cool. And mining operations require large volumes for mineral processing. Yet despite the rising competition, regulatory systems haven’t kept pace, the report finds.
“All of these sectors are simultaneously converging on the region,” Helena Volzer, senior source water policy manager at the Alliance for the Great Lakes, and author of the report, tells Sentient. “And so that’s why we’re sort of raising this concern that states need to be doing this kind of regional demand planning to figure out where water is available and where it’s not.”

Groundwater plays a critical role in the Great Lakes ecosystem, feeding streams, wetlands, and even the lakes themselves. An estimated 20 to 40 percent of the Great Lakes’ water budget originates from groundwater, and 40 to 75 percent of residents in the Great Lakes states rely on it for drinking.

The report notes that mining uses a significant amount of water, particularly for critical minerals like lithium, cobalt, copper, and nickel used in clean energy technologies and data center equipment.

While mining is important, this analysis focuses on comparing data centers’ water demands — and lack of transparency — to agriculture, whose water use remains deeply embedded in both regional and global water systems.

So how do the water footprints of the agriculture and data center sectors compare in reality?

Data Centers Are Quietly Making a Grab for Water

Data centers, especially those powering artificial intelligence, have captured headlines for their growing thirst for water. The Alliance for the Great Lakes report states that a hyperscale data center can consume more than 365 million gallons annually, which is roughly the same as the water use of 12,000 Americans over the same period. In 2023, U.S. data centers consumed an estimated 17.4 billion gallons of water — a figure expected to double by 2028 as cloud computing and AI demands surge. Yet despite this growing water demand, transparency around water usage remains elusive.

“We really don’t know what the best water conserving practices are inside of a data center,” Volzer says. Cooling methods vary widely, she adds, which is why one of the things the report calls for is more conservation and efficiency standards.

Cooling accounts for the majority of data centers’ water use. When evaporative cooling systems are used, most of the water evaporates and is not returned to its watershed. If powered by fossil fuels or nuclear energy, even more water is needed to cool their power generation systems, compounding the loss of the finite resource.

Less than one-third of data centers track their water use consistently, largely because no regulations require it. Bloomberg News found that two-thirds of new data centers are being built in areas experiencing high water stress, raising further concerns. And two Great Lakes states, Illinois and Ohio, rank among the top five in the nation for the number of data centers.

But data centers aren’t the only sector with a heavy water footprint, the AGL report notes. Agriculture is a long-established user, particularly for growing feed crops for livestock.

How Agriculture Shapes Water Demand

Agriculture, the backbone of food systems around the world, is also the largest global consumer of freshwater, accounting for nearly 70 percent of global water withdrawals, according to the U.N. World Water Development Report.

In the Great Lakes region, irrigation accounts for 23 percent of water use, compared with 25 percent for industry and 33 percent for public supply. Irrigation was a major driver of U.S. freshwater use, accounting for 42 percent of total withdrawals as of 2015.

Irrigated crops in the U.S. have shifted over time. In 1964, the leading irrigated crops were cotton, hay and alfalfa. Today, corn (for grain), soybeans, alfalfa and hay dominate. These crops are primarily grown for livestock, not for human consumption. According to the USDA, this shift reflects “expanding market demand for corn and soybeans as livestock feed and as a source for biofuel.”

This high water use for irrigating crops diverts millions of gallons per day from the lakes and aquifers, significantly impacting ecosystems, wildlife and long-term water availability. Limited water also restricts the acreage farmers can irrigate.

“With irrigation of crops, that [water use] can fluctuate, especially during peak summer months,” Volzer noted. “And we’re seeing irrigation being used now in places where it historically hadn’t.” For example, Michigan is experiencing longer drought conditions that are driving increased irrigation demand.

“Agriculture, for example, doesn’t use that much water in the aggregate, but it has a high consumptive use component to the water that they are using. A lot of it evaporates in the process.” Volzer says.

Evaporation becomes especially critical when multiple sectors pull from the same water sources at the same time. For example, during peak summer heat, farms need more water for crops, while data centers need more water to stay cool, compounding demand on already strained systems.

How to Safeguard the Great Lakes’ Water Supply

Both agriculture and data centers compete for the Great Lakes’ precious freshwater, but their impacts differ. Data centers are rapidly expanding, with little regulatory oversight on water use. Agriculture, by contrast, remains deeply entrenched in water consumption through feed crop irrigation and livestock production.

Training large AI models, for instance, requires intensive computing power, which translates into significant water use for electricity generation. The Washington Post found that training one version of Chat GPT-3 consumed as much water as producing 100 pounds of beef, nearly double what the average American eats in a year.

While AI training is energy and water-intensive, its overall water footprint remains small compared to agriculture’s entrenched demands. It takes roughly 660 gallons of water to produce a single hamburger. In comparison, ChatGPT consumes about 0.000085 gallons of water per response, less than one-fifteenth of a teaspoon.

Scaling back industrial meat production, particularly beef, could dramatically reduce the demand for feed crops and the water needed to grow them. Such a shift would ease pressure on the Great Lakes’ fragile water systems and reduce greenhouse gas emissions from livestock, which account for about 14.5% of global emissions, helping to mitigate climate change and the worsening droughts it brings.

The AGL report recommends requiring all large-scale water users to report their usage. It also calls for public disclosure of proposed water and energy use, the removal of data center-specific tax incentives, and updated conservation standards tailored for modern water stress. Strengthening groundwater protections and funding comprehensive mapping of aquifer recharge rates are also key recommendations.

“Regional demand planning needs to be happening,” Volzer says. “But at the same time, we need more transparency around water and energy use, so public disclosure requirements around how much water is being proposed to be used.”

Without action, both industries risk accelerating the depletion of the Great Lakes’ most valuable and finite resource.

This piece has been updated to correct a name misspelling.