LG Innotek is partnering with Nvidia to develop sensing solutions that help robots see their surroundings and manipulate objects, the South Korean electronics component maker announced on Oct. 22. The collaboration focuses on physical AI, which describes AI systems that interact with the physical world through machines like robots and autonomous vehicles. At Nvidia's GTC Berlin conference this week, LG Innotek demonstrated a 3-megapixel camera module and a multimodal vision system built around Nvidia's Holoscan Sensor Bridge technology.
The multimodal vision system combines multiple cameras and sensors to provide AI systems with a more complete view of their surroundings, drawing on LG Innotek's experience in optical design and sensing technologies that span cameras, LiDAR and radar. Nvidia's Holoscan Sensor Bridge technology is designed to transfer sensor data efficiently to AI computing systems while cutting back on intermediate CPU processing, which can help lower data-transfer delays and allow robots to interpret their environments and respond more quickly. The companies plan to extend their work together to humanoid robot vision systems and robotic hands, with next-generation solutions expected to be showcased at CES and GTC in 2027. The partnership fits LG Innotek's push to expand beyond its established mobile-device component business and build new revenue streams in robotics and autonomous driving, as the company is already developing robotic components with Boston Dynamics and has begun mass-producing camera modules for a North American robotics customer.
LG Innotek CEO Moon Hyuk-soo said the Nvidia collaboration offers a chance to expand the company's integrated sensing technologies globally, according to BigGo Finance. He also described the company's ambition to become a physical AI solutions provider spanning sensing, communication modules and semiconductor substrates. For robotics manufacturers and systems integrators, the collaboration could eventually expand the range of sensing components designed to work with Nvidia's AI computing platforms, which may create opportunities to develop more integrated robotics solutions, though the companies haven't announced specific channel programs, pricing or commercial availability for the new systems.
The challenge for robot makers isn't simply installing better cameras, but turning incoming visual information into reliable actions quickly enough to navigate spaces, recognize objects and manipulate items without constant human intervention. Connecting sensors directly to AI computing systems could help reduce processing delays, while combining cameras with LiDAR and radar could give robots additional information about distance and their surroundings, though the results will depend on sensor quality, software and how well the components work together. For businesses deploying robots, better integrated sensing could eventually mean machines that recognize objects more accurately, react faster to obstacles and perform handling tasks with fewer errors — in warehouses, for example, improved perception could help robots identify packages and navigate around workers or equipment. The partnership's commercial impact will depend on whether LG Innotek and Nvidia can turn their sensing technology into products that robot manufacturers can deploy at scale, with the next milestones being the planned 2027 demonstrations, followed by evidence of real-world performance, customer adoption and commercial availability. The companies will need to prove their technology works reliably in varied environments before channel partners and customers commit to large-scale adoption. Whether this collaboration reshapes how robotics systems integrate perception hardware will hinge on execution timelines and the willingness of manufacturers to standardize around a shared platform architecture.