The last decade of AI progress was built on text, but the frontier has shifted toward data from the physical world. Video, LiDAR point clouds, sensor streams, and other high-dimensional data now drive systems that perceive, reason, and act in physical space. This report, based on a 2026 survey of more than 700 professionals, documents how teams actually build physical AI today. It finds that data problems cause the majority of model failures, and that curating data matters more than chasing larger architectures. Annotation remains costly and wasteful, because teams often label everything and then discard much of it before production. The findings show why data work, not data collection, separates teams that ship from teams that stall.
IEEE Spectrum Magazine, the flagship publication of the IEEE, explores the development, applications and implications of new technologies. It anticipates trends in engineering, science, and technology, and provides a forum for understanding, discussion and leadership in these areas.
Do you have a query about the content on the hub or are having issues with access?
Source link
