GIST and KENTECH develop brain-inspired AI memristor

Researchers from the Gwangju Institute of Science and Technology (GIST) and Korea Institute of Energy Technology (KENTECH) have developed a next-generation volatile memristor designed to mimic biological neurons. By using the Marangoni effect to achieve self-assembly, the team successfully arranged cobalt oxide (Co3O4) single-crystal nanoparticles into a regular checkerboard pattern with precise 3nm nanogaps.

This structural design addresses a common issue in oxide-based memristors where inconsistent conductive paths lead to unstable performance. The controlled nanogaps ensure a consistent path for electrical current, allowing the device to operate with high stability and low voltage. This breakthrough offers a new direction for neuromorphic computing and Internet of Things (IoT) technologies, which require high-performance electronic components that consume minimal power.