Nvidia is enhancing collaborations with local component manufacturers LG Innotek and Samsung Electro-Mechanics, bolstering its supply chain for physical artificial intelligence (AI) applications and upcoming AI infrastructure.
LG Innotek revealed a range of sensing solutions featuring Nvidia's newest Holoscan Sensor Bridge (HSB) technology at the annual GPU Technology Conference (GTC) held in Berlin.
Also Read: Aldar Seals Deal for Premium Sharjah Residential Units
The offerings consist of 3-megapixel camera modules and a multimodal vision system that integrates various cameras and sensors, aimed at uses in physical AI and robotics.
Nvidia's HSB technology enables sensor data to be transmitted directly to AI computers, avoiding the central processing unit and possibly accelerating the way robots and autonomous vehicles analyze information and reach decisions. The collaboration demonstrates increasing demand for sophisticated sensing technologies as AI evolves from software and data centers into tangible environments, where machines need to sense and react to their surroundings instantly.
LG Innotek announced its intention to enhance its partnership with Nvidia regarding HSB-based sensing technologies for vision systems in humanoid robots and robotic hands. The firms aim to create sensing solutions that are compatible with Nvidia's forthcoming platforms and present them at CES and GTC in the coming year.
“This collaboration is a valuable opportunity to bring LG Innotek’s differentiated integrated sensing solutions to global markets,” company CEO Moon Hyuk-soo says.
Also Read: Singapore, Google Develop On-Device AI Tool to Detect Scams
“As a total solution provider for physical AI, encompassing not only sensing but also communication modules and semiconductor substrates, we will move quickly to establish a leading position in the physical AI solutions market.”
In a separate announcement, Samsung Electro-Mechanics revealed a strategic partnership with Nvidia aimed at increasing the production of advanced semiconductor package substrates for AI chips and data center network switches. Nvidia is said to be investing around 2.6 trillion won ($1.97 billion) in the expansion, but Samsung Electro-Mechanics has not verified this figure.
The firms will collaboratively finance new manufacturing capacity, with Samsung Electro-Mechanics scheduled to start providing Nvidia with the substrates in 2026 under a multi-year contract.
Also Read: EAM Jaishankar Urges Urgent Multilateral Reform, Cites India-UAE Ties
The firm plans to set up specialized production lines at its sites in Sejong and Busan to produce substrates for Nvidia's computing chips and high-speed network switches connecting chips in data centers. The firm projects that its yearly substrate deliveries to Nvidia will increase by an average of double digits starting in 2026. By 2028, orders related to Nvidia are projected to represent a significant portion of the overall package substrate production capability of the company.
Advanced substrates are gaining significance as AI chips become more intricate and demand enhanced computing power, stricter power management, and more effective data transmission. The firm stated that its manufacturing methods encompass fine-line circuit production and the integration of passive components, including multilayer ceramic capacitors and silicon capacitors, into substrates.