Researchers from the University of Cambridge have developed a new chip technology inspired by the human brain that can cut AI energy usage by up to 70%. This breakthrough centers on the memristor, a component that mimics biological neurons.
Unlike traditional computing architectures that constantly move data between the processor and memory, memristors store and process information in the same location. This efficiency eliminates the energy-heavy data transfers responsible for high consumption in current AI systems.
This advancement in neuromorphic computing offers a sustainable path forward for large-scale artificial intelligence, significantly reducing the carbon footprint of future data centers.
