homehome Home chatchat Notifications


New Liquid Metal Tech Outperforms Existing Solutions by 72%

This new material promises to cut data center cooling costs and emissions.

Tibi Puiu
November 12, 2024 @ 7:35 pm

share Share

Credit: University of Texas at Austin.

Data centers are indispensable in our digital age, where the demand for cloud computing, artificial intelligence, and streaming services continues to surge. But these server farms, packed with powerful processors, are also energy-hungry beasts. They require vast amounts of electricity to keep cool, driving up operational costs and contributing to carbon emissions. Now, a team of researchers from the University of Texas at Austin have come forward with a potential solution.

Researchers developed a new “thermal interface material” (TIM) designed to dissipate heat more efficiently than anything currently available on the market. This material, made by blending a liquid metal alloy called Galinstan with ceramic aluminum nitride, outperforms the best commercial liquid metal cooling products by up to 72%.

Not your typical CPU thermal paste

Data centers, which power everything from cloud storage to AI models, generate enormous heat. Cooling these facilities accounts for roughly 40% of their total energy consumption. Globally, that translates to around 8 terawatt-hours of energy use annually—enough to power entire cities. As AI and machine learning applications expand, the pressure on data centers is only set to grow. Goldman Sachs recently predicted that data center power demand could surge by 160% by 2030, with AI alone potentially adding 200 terawatt-hours per year in power consumption over the next decade.

Schematic of essential components in power devices. Credit: The University of Texas at Austin.

This is where the new TIM developed at the University of Texas comes in. The researchers achieved a remarkable feat by combining the liquid metal with ceramic particles using a process known as mechanochemistry. This method creates controlled, gradient interfaces within the material, allowing heat to flow through it more efficiently than ever before. The result is a TIM capable of dissipating an impressive 2,760 watts of heat from an area as small as 16 square centimeters.

“This breakthrough brings us closer to achieving the ideal performance predicted by theory,” said Kai Wu, the lead author of the study. “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.”

Impact Beyond the Lab

The new material could drastically reduce the need for energy-intensive fans and pumps in data centers, cutting their energy usage for cooling by up to 65%. If deployed industry-wide, this technology could lower cooling energy demands by 13%, reducing overall data center power consumption by at least 5%.

Guihua Yu, a professor in the Cockrell School of Engineering and Texas Materials Institute, emphasized the urgency of finding new solutions. “The power consumption of cooling infrastructure for energy-intensive data centers and other large electronic systems is skyrocketing,” Yu noted. “That trend isn’t dissipating anytime soon, so it’s critical to develop new ways, like the material we’ve created.”

But the journey from the lab to the real world remains a challenge. So far, the team has only tested the material on a small scale in controlled lab environments. The next step involves scaling up production and conducting field trials with data center partners. If successful, this new cooling technology could soon become a staple in the industry, allowing server farms to run more processors in the same space without overheating.

While it may take years before this technology is available for consumer use, the implications for large-scale data operations are clear. By making data centers more efficient, it could help curb the world’s ever-expanding environmental digital footprint.

The findings appeared in the journal Nature Nanotechnology.

share Share

Ants outperform humans at group puzzle-solving activity

Ants may have tiny brains, but when it comes to teamwork, they pack a mighty punch.

Geneticists have finally solved the mystery of Garfield’s orange coat

Two new studies have revealed why some cats are orange – an enduring enigma of genetics, until now.

What did Roman wine taste like? It was 'spicy' and had an orange color

The secrets of ancient Roman wine are being uncorked by modern science.

The Science Behind Why Labradors Are Always Hungry

Labrador owners can finally stop feeling guilty for overfeeding.

Leopards have unique roars, and AI can identify them

They say you can identify the leopard by its spots, but as it turns out, you can also identify it through its unique roar. Leopards, notoriously difficult to monitor due to their elusive nature, could soon be tracked using passive acoustic recorders paired with AI-based analysis. Leopard populations face significant challenges, with their ranges shrinking […]

Cosmic fireworks: zombie star explodes, creating massive filament structures

This incredible image captures the ghost of a supernova 100 light-years across.

3D-printed 'ghost guns', like the one Luigi Mangione allegedly used to kill a health care CEO, surge in popularity as law enforcement struggles to keep up

The use of 3D-printed guns in criminal and violent activities is likely to continue to increase. And governments and police will continue to have trouble regulating them.

The Billion-Year Journey That Shaped the Universe We Know Today

The revolutionary James Webb Space Telescope and next-gen radio telescopes are probing what’s known as the epoch of reionization. It holds clues to the first stars and galaxies, and perhaps the nature of dark matter.

Some Cultures Have No Words for Numbers Beyond 'Three'. Here's What They Can Teach Us

Can you imagine a world without numbers? For many people, that's their reality.

These Revolutionary Maps Are Revealing Earth's Geological Secrets

This work paves the way for more precise and comprehensive geological models