Oregon State Builds Brain-Inspired Sensor to Cut Edge AI Power Use

Conventional AI vision hardware separates sensing, memory and computation, forcing image data to shuttle among components and raising power use and latency. Researchers at Oregon State University are pursuing a neuromorphic, in-sensor approach that lets hardware retain, weight and discard signals at the point of capture. The concept matters for robots, drones, security cameras and other edge AI systems, where battery life, response time and limited connectivity make local, energy-efficient processing particularly valuable.

Oregon State announced the prototype on June 16, 2026, after the research was published in Advanced Functional Materials with support from the U.S. National Science Foundation; no funding amount was disclosed. Led by electrical engineering and computer science professor Larry Cheng, the team built a 4-by-4-pixel phototransistor array roughly the size of a USB stick. It pairs an indium gallium zinc oxide, or IGZO, channel with an organic photosensitive layer. Gate voltage shifts trapped charges, allowing optical memories to persist for hours or longer or fade faster, though the technology remains at the device-demonstration stage.