For five weeks, students Mason Hernandez ‘29 (Mechanical Engineering), Marwa Zeroual ‘29 (Mechanical Engineering) and David Cacsire ‘29 (Mechanical Engineering) worked with faculty mentor, Associate Professor Ping-Chuan Wang as part of the University’s AC2 Summer Research Program, an experiential learning opportunity that immerses New Paltz students from STEM-related disciplines into real-world research experiences.

Professor Wang’s years of scholarship on the issue of cooling down digital devices and power centers made him the perfect guide for this team of young engineers to develop an innovative, potential solution to the energy and water consumption issues increasingly associated with artificial intelligence data centers.

“With AI expanding at a fast rate, data centers continue to utilize a large supply of water in order to cool down their electronic components,” said Zeroual. “By working together to create a system that conserves liquid consumption, we would be able to reduce both energy costs and environmental hazards while using AI.”

Together, the student-faculty team devised a sustainable prototype for an automated device that could cool down computer chips needed to power data centers while minimizing impacts on water and energy.

“One unique thing about this AC2 Summer Research Program is that you have multiple students with different technical backgrounds and grade levels,” said Wang. “In this case, we had three first-year students, which helped bring a whole different energy because they were eager to learn and contribute new ideas.”

For example, Hernandez suggested that the student-faculty team make use of the University’s Hudson Valley Additive Manufacturing Center to build their prototype from scratch through 3D printing.

“I find 3D printing to be quite exciting in terms of what it can produce, and I felt it would be most appropriate to work on campus within this facility to create what we needed in a short amount of time,” he said.

The group’s early efficiency studies suggest that at scale, their device – a 3D- printed cold plate structure with a microcontroller that would open and close depending on the temperature of the chip – has the potential to save up to 85% of a data center location’s water supply.

“What we envisioned would cool down to a set temperature, and then it would cut off the stream of water from flowing,” said Hernandez.