For our series, A Fordham Focus on AI, we’re speaking with faculty experts about the impact of this rapidly evolving technology and what to expect next. We sat down with Kim Diamond, an adjunct professor at Fordham Law whose expertise spans science, technology, and energy policy. This fall, she added the subject of data centers to a class on energy law she’s taught for the past decade. As AI systems require massive data centers to operate, those facilities demand enormous amounts of electricity—and water to cool both the servers and the power plants that run them.
Why are you worried about how data centers will affect our freshwater supply?
You’re going to need a vast amount of water to meet the energy needs of all the planned data centers. Water is used to cool facilities like nuclear power plants. Similarly, water is used on-site in data centers to keep computer servers from overheating. A report from Lawrence Berkeley National Laboratory found that in 2023, data centers used about 211 billion gallons of water—enough to fill approximately 320,000 Olympic-sized pools.
And the number of these centers is growing quickly. Data center construction spending nearly doubled between 2022 and 2024, while electricity use increased from 250 to 400 terawatt-hours between 2020 and 2024. This has the potential to drastically affect the health of the aquifers people depend on.
So what legal tools exist to address this kind of environmental strain?
When industries put pressure on shared environmental resources, the law typically responds in one of two ways: through tort law or through the public trust doctrine.
Tort law allows individuals or communities to sue companies for harm. But these cases are often difficult to win in environmental contexts. For example, a Native Alaskan village sued ExxonMobil and others in 2008, arguing that these companies’ contributions to climate change made their land uninhabitable. The case was unsuccessful, in part because courts found it difficult to connect specific environmental harm directly to these particular companies.
That kind of outcome has led scholars and environmental advocates to explore another approach: the public trust doctrine. Under this theory, governments have a duty to protect essential natural resources—such as air and water—for the public’s benefit.
So far, no public trust case has targeted AI data centers’ water use. But some scholars are exploring whether novel constitutional arguments could expand the conversation.
Consider the Fifth Amendment. It protects against being “deprived of life” without due process of law. The 14th Amendment, which applies to the states, says the same. So what is fair and due to us as citizens, and when should we have a fair say? Is our future access to water a right? That’s important to consider, because if an aquifer is completely drained by data center-driven demand, it’s not like we can replenish it.
We haven’t seen a court address that argument in the context of AI data centers. But if such a claim were successful, it could shape environmental oversight nationwide and encourage the development of cleaner, more water-efficient systems.
What can people who aren’t lawyers do?
There’s a line from the Rush song, “Freewill,” that I really like: “If you choose not to decide, you still have made a choice.” If we decide to do nothing, that is still a choice. There are ways people can contribute, whether it’s attending public hearings, weighing in with comments during hearings, or reaching out to their congressperson.
You’ve said you’re in favor of leaders tackling the issue through new legislation. Why?
The legislative approach avoids many of the risks associated with litigation, takes public input into account, and can be enacted more quickly and less expensively.
We’re beginning to see examples of this approach at the state level. In 2025, Minnesota passed a law requiring developers of large AI data centers to undergo a review process before approval—including a pre-application evaluation for projects expected to use more than 100 million gallons of water per year.
As we grapple with the impact of data centers on the environment, what gives you hope?
Because the scale-up is so large and widespread, people are starting to connect the dots about the environmental impacts of data centers. Also, we know technological innovation is always possible, so it’s just a matter of how much we want to accelerate improvement in water efficiency now, so we can still get the bang for the buck from what data centers are doing while minimizing their water intake.
There’s an old saying among the Iroquois League, known as the Haudenosaunee: when you look to the future, you should couch your decisions in terms of how they will benefit everyone seven generations from now. Applying that principle to AI development means thinking not only about innovation, but about the resources future communities will need.
Learn more about AI for the greater good at Fordham.
