Scientists at the University of Texas are getting ahead of one of the biggest challenges of the artificial intelligence revolution.
That is how to more efficiently cool down massive data centers, as the demands on them rapidly increase to power programs like ChatGPT.
Nature Magazine published the researchers' encouraging findings using a bespoke "thermal interface material" to cut the energy required for cooling pumps to lower the temperature of electronic devices.
In a news release, the scientists say their tech created up to 65% in energy savings for the pumps. Considering the Department of Energy said that data center cooling accounts for up to 40% of their energy usage, that could be a big deal.
The Texas researchers said that if adopted on a wider scale, the thermal material can cut cooling energy usage by 13% and overall usage by 5%.
The demand for data centers is currently exploding, with Goldman Sachs projecting a 160% increase in demand by 2030. AI is a huge part of that. A ChatGPT inquiry requires almost 10 times as much electricity as a Google one.
"The power consumption of cooling infrastructure for energy-intensive data centers and other large electronic systems is skyrocketing," said Guihua Yu, one of the researchers.
It's why the UT scientists are among many efforts to optimize cooling like a proposed University of Missouri solution. Data centers require an enormous amount of energy and are often fueled by dirty energy that is warming the planet with disastrous results.
Finding a way to increase efficiencies in the cooling process will be key to stemming some of the biggest potential problems with the soaring demand.
🗣️ Do you think our power grid needs to be upgraded?
🔘 Definitely 👍
🔘 Only in some states ☝️
🔘 Not really 👎
🔘 I'm not sure 🤷
🗳️ Click your choice to see results and speak your mind
Powering data centers with clean energy sources, like Google is doing in an Arizona location, will also be critical. Similarly, efforts to make AI itself more efficient can play a role, as scientists at the University of Minnesota, Twin Cities, demonstrated in recent research.
According to the news release, thermal interface materials are specifically designed to dissipate the heat from technological devices. In the past, though, the theoretical cooling abilities didn't translate well in real-life applications.
For the study, the scientists used mechanochemistry to mix liquid metal and aluminum nitride. The resulting thermal interface material delivered when tested on small lab-scale devices.
"This breakthrough brings us closer to achieving the ideal performance predicted by theory, enabling more sustainable cooling solutions for high-power electronics," said Kai Wu, a lead author of the study.
Next for the University of Texas team will be to scale up the size of their materials, so they can test larger versions with data center partners. The team is excited about the possibilities of their findings to make a major impact in cooling tech globally.
"Our material can enable sustainable cooling in energy-intensive applications, from data centers to aerospace, paving the way for more efficient and eco-friendly technologies," Wu concluded.
Join our free newsletter for weekly updates on the latest innovations improving our lives and shaping our future, and don't miss this cool list of easy ways to help yourself while helping the planet.
