Scientists develop ultrafast magnetic-field pulses memory system to slash AI data center energy use

Researchers at the University of Edinburgh have developed a new method for magnetic memory that could significantly reduce energy consumption in AI data centers. By utilizing ultrafast magnetic-field pulses instead of traditional energy-intensive switching techniques, the team claims they can cut energy usage by up to 100 times. The study, published in the journal Advanced Materials, suggests that this approach could help mitigate the environmental impact of the rapidly growing AI industry, which currently faces criticism for its high electricity demands. While the research is currently theoretical, the authors believe the framework could be adapted for electrical currents and ultrafast laser pulses, potentially revolutionizing future data storage technologies. Although practical implementation remains a long-term goal, the breakthrough offers a promising path toward more sustainable computing infrastructure and a reduction in the carbon footprint associated with global data processing.