Finans Nyheter: 29 juli 2025 - 21 september 2025

December 15, 2025, 9:18 PM EST. Researchers from the University of Kent led by Frank Zhigang Wang demonstrate a meminductor-an inductor whose inductance L(q) changes with the history of charge, providing a memory function unlike traditional components. The team shows that this charge-dependent inductance can reproduce the memorizing, timing, and anticipatory behaviors observed in amoebae, offering a new path beyond the memristor for neural-inspired architectures. By linking memory to the time constant of circuits, meminductors open possibilities for neuromorphic computing and deep learning applications, potentially improving energy efficiency and scalability. This work challenges the dominance of von Neumann architectures, suggesting integrated memory-processing devices can better emulate primitive intelligence and biological dynamics in AI systems.