Student-Led Panel Confronts AI’s Growing Need for Energy

Artificial intelligence is often described as invisible, but at Stony Brook University, a panel on sustainability pointed instead to the physical systems that make AI possible. Behind every query or prompt, speakers said, are significant demands on energy, water and infrastructure.

The panel "AI and Sustainability: Can AI's Energy Demands Be Sustained?" drew an audience of students, faculty, staff and community members to the Sidney Gelber Auditorium April 27 to take a closer look at how the rapid growth of AI is affecting the environment.

The event was created and organized by students in the School of Marine and Atmospheric Sciences (SoMAS) through Senior Lecturer Tara Rider's SUS 401 capstone course, with students introducing the panel and serving as moderators.

"AI does not exist without the foundational elements that are concrete, steel, water and electricity," a student speaker said at the start of the event.

Panelists included Hendrik Hamann, professor of Atmospheric Sciences and AI chief scientist at Brookhaven National Laboratory (BNL); Nina Kshetry, president of Ensaras, a provider of AI solutions for wastewater; and Aaron Miller, eastern regional manager for SHARC Energy, a Canadian clean-tech company that specializes in wastewater energy transfer. They approached the issue from different angles, but each addressed how to balance rapid AI growth with the resources needed to sustain it.

"AI is computing," Hamann began. "What is the basis of AI? What is AI? AI is computing computation."

Concerns about rising energy demand are not new, he added, pointing to earlier predictions about data center growth that never fully materialized. "It didn't happen because there was a wake up call for everyone and a lot of innovation happened," he said.

"I am actually rather optimistic," Hamann added, noting that continued advances in hardware and system design could help slow the trajectory again.

Miller, whose work focuses on energy recovery systems, put it more bluntly. "Power and water consumption are massive drivers on the market," he said, adding that his experience has shown "the negative possible impacts of AI on the environment and on the grid." But he also pushed back on the idea that those impacts are fixed. Data centers, he said, can be rethought as energy hubs, capturing excess heat and putting it to use in nearby communities instead of letting it go to waste.

"The timeline is now," Miller said. "Scale is key."

Kshetry discussed how connected these issues are, describing what she called a growing "nexus" between water, energy and AI. "Everything uses each other," she said. Her work focuses on improving efficiency in wastewater and industrial systems, something she said could help reduce the environmental impact of data centers, particularly when it comes to water use.

"There are ways that we can use wastewater to help mitigate the impact data centers have on local community fresh water sources," she said.

Still, even with those kinds of solutions, panelists made it clear that technology alone will not solve the problem. Policy, economics and community impact all shape how quickly changes happen and who ultimately benefits. Hamann noted that for many communities, data centers are still seen as more of a burden than an asset.

"This is an ugly building, it increases my electricity cost, it uses up my water," he said.

Miller pointed to the financial reality behind those concerns. "Nothing's going to work in America if people can't make money on it," he said.

As the conversation turned to careers, panelists pointed to growing demand not just for technical expertise, but for people who understand how these systems work in practice. "Domain sciences will excel," Hamann said, emphasizing the importance of subject matter expertise. Kshetry pointed to opportunities in areas like AI orchestration and infrastructure systems, but made it clear: "You still need to know how to fix them," she said.

Questions from students and community members turned to workforce changes and what they might mean for future careers. One audience member, a trades apprentice, shared that his brother had worked in computer science for years before being laid off and struggling to find new work.

Miller said students will need to stay adaptable as the field continues to shift. "There's going to be a ton of really cool spaces where you're going to be able to carve out your own niche," Miller added.

Panelists agreed that meeting the demands of AI will depend on how well technology, infrastructure and policy work together.

"We have what we need," Miller said. "It's just a matter of allocating it."

- Beth Squire