U of M researchers launch startup aiming to reduce AI energy consumption

Professor Jian-Ping Wang and his team at the University of Minnesota are pioneering a new computing technology called Computational Random Access Memory (CRAM), designed to dramatically reduce the energy demands of artificial intelligence systems. Traditional computers consume significant energy by constantly moving data between memory and processors, a process CRAM aims to eliminate by enabling data processing directly within memory. Wang's research, which dates back over two decades, is now being accelerated by advancements in materials and device design. To bring this innovation to market, Professor Wang recently founded BesiMax, a startup focused on developing silicon-based chips that transform memory into processing units. The company aims to integrate these chips into major data centers, with a prototype expected within two years. This initiative addresses growing concerns about the escalating energy and water consumption of data centers globally, with BesiMax estimating a potential 99% reduction in AI energy use.