Lyte raises $107M to power perception layer for physical AI robots

Lyte is going after one of the hardest and least glamorous parts of physical AI: robust, standardized perception. By bundling depth sensing, RGB and motion awareness into a single hardware‑software system, it’s trying to remove a major integration tax that every robotics and autonomous vehicle team currently pays. The founders’ track records on Kinect and Face ID make this more than a speculative bet—they’ve already shipped perception systems at billion‑device scale.

In the broader race to AGI, the story here is that large‑scale general intelligence will almost certainly extend into the physical world, not remain trapped in chat windows. That requires perception stacks that are consistent, calibratable and easy to deploy across many robot form factors. If Lyte becomes a default choice for that layer, it could play a role analogous to CUDA or ROS—part of the implicit infrastructure on which embodied intelligence advances.

The size of the round and caliber of investors suggest serious conviction that the bottleneck in robotics is shifting from algorithms to deployment complexity and sensing reliability. For AGI‑adjacent investors, it’s a reminder that the path to general physical intelligence may run through companies that look more like “boring infrastructure” than sci‑fi robots.