Scientists discover a new way AI data centers can slash their energy usage (www.techradar.com)

Researchers at the University of Edinburgh have made a significant breakthrough in reducing the energy consumption of AI data centers by employing "ultrafast magnetic-field pulses" in RAM and storage systems. This innovative technique offers a potential energy savings of up to 100 times compared to traditional magnetic memory switching methods, which are currently energy-intensive. As the demand for AI continues to surge, the associated energy costs and environmental impact are becoming increasingly pressing issues, with predictions of rising contributions to global electricity consumption and carbon emissions. Although the findings, published in the journal Advanced Materials, are still theoretical, the implications are promising. The researchers believe that their framework could be adapted for other advanced technologies, such as electrical currents and ultrafast laser pulses used in future data storage solutions. While practical implementations may take time to materialize, this discovery holds the potential to transform the energy landscape for data centers, encouraging more sustainable practices in the rapidly evolving AI and computing industries.