In a bid to enhance the efficiency of AI data centres, researchers in Switzerland are pioneering new cooling technologies aimed at reducing both energy and water usage. With the rise of artificial intelligence driving increased demand, scientists at the Swiss Federal Institute of Technology in Lausanne (EPFL) and the start-up Corintis are collaboratively working on sophisticated liquid-cooling systems. These systems are specifically engineered to align with the thermal properties of individual computer chips, potentially cutting cooling energy consumption by approximately 20% while utilizing less water than traditional cooling methods.
A significant aspect of this project is the innovative approach to managing waste heat produced by AI servers. Instead of allowing this heat to dissipate into the atmosphere, the researchers have formulated plans to repurpose it. During winter months, the waste heat could be redirected into district heating systems, whereas in warmer seasons, a portion of this energy could be converted back into electricity, thus enhancing overall energy efficiency.
As AI processors are known to generate substantially more heat compared to conventional computing systems, experts emphasize the growing necessity for liquid cooling solutions. This development is seen as vital in addressing the thermal challenges posed by advanced AI technologies. However, specialists also caution that while improved cooling technologies are crucial, they alone will not mitigate the environmental impact of AI. The ever-increasing demand for AI services globally continues to escalate electricity and water consumption.
The Swiss researchers’ efforts reflect a broader push within the tech industry to adopt more sustainable practices amidst the rapid expansion of AI capabilities. By focusing on both reducing resource consumption and creatively reusing waste heat, these advancements in cooling technology represent a crucial step towards more environmentally responsible data centre operations.