Why video is the first frontier as AI learns to read the physical world
Lumana, founded by ex-Intel computer vision leaders, is making existing video feeds from 562 million deployed surveillance cameras intelligent and searchable. The company processes over a billion images daily across 50,000+ cameras using deep learning to enable physical AI applications.
The Atomic take
Lumana's bet is that the cheapest path to physical AI isn't new hardware but the 562 million cameras already bolted to walls, turning passive surveillance feeds into searchable data at a scale of over a billion images a day. That matters because the bottleneck for robotics and physical AI has always been perception, and retrofitting deployed infrastructure sidesteps the cost and delay of installing new sensors. Watch whether Lumana can move beyond search and alerts into feeding this visual understanding directly to robots and automated systems, and how it handles the privacy backlash that comes with making every existing camera intelligent.
Read the full story at TNW →Know where humanity is headed.
Five forces, one investment thesis a day. Free, every weekday morning.
Get The Atomic Five