In-memory computing, which seeks to reduce power consumption by processing data where it's saved, is seeing a boost in interest fueled by the AI energy crisis (Jane Lanhee Lee/Bloomberg)