cambridge

Cambridge's brain-mimicking chip material cuts AI power draw a millionfold

Cambridge researchers built hafnium oxide memristors with switching currents a million times lower than conventional devices, a material breakthrough that could cut AI hardware energy use by 70%.