LG Innotek is extending its camera and sensing business into robotics with new vision hardware designed to work with NVIDIA’s Holoscan Sensor Bridge. The South Korean components manufacturer disclosed the development on October 22, following a demonstration at NVIDIA GTC Berlin on October 20.
The lineup includes a three-megapixel camera module and a multimodal vision system that combines information from different types of sensors. Both use NVIDIA’s Holoscan Sensor Bridge technology, which moves sensor data directly to an AI computer rather than routing it through a conventional CPU. The architecture is intended to reduce the delays between a machine observing its surroundings and processing what it sees, an important consideration for robots that must react while moving or manipulating objects.
LG Innotek has extensive experience supplying optical components for mobile devices and autonomous vehicles. Its robotics work is now drawing on that expertise, including the integration of cameras with other sensing technologies such as lidar and radar. The company said it has already begun mass production of camera modules for a North American robotics customer, although it did not identify the buyer or disclose shipment volumes.
The NVIDIA collaboration will also extend to vision systems for humanoid robots and sensing technology for robotic hands. LG Innotek plans to show further products compatible with NVIDIA’s next-generation platforms at CES and GTC events in 2027. It has not provided pricing, production dates or customer commitments for the newly demonstrated Holoscan-compatible lineup.
The October announcement concerns sensing components and development plans, not the launch of a new humanoid robot. For manufacturers building machines that need to perceive their environments and handle objects, the performance and integration of cameras and other sensors are practical engineering constraints. LG Innotek is positioning its established component business to serve that market.
