Endlight Partners with NVIDIA to Expand Physical AI Data Infrastructure with TRINIX

By StartupKorea Business Desk | Jun 08, 2026

Endlight Accelerates Entry into Physical AI Market

Endlight, a startup specializing in 3D AI technology, has announced a strategic partnership with NVIDIA to enhance its presence in the Physical AI sector. The collaboration centers around Endlight's TRINIX platform, which integrates with NVIDIA's Omniverse and Isaac Sim to bolster data infrastructure for global manufacturing and robotics.

Innovative Data Solutions for Manufacturing

TRINIX is designed to facilitate the automatic generation and transformation of data assets, enabling the seamless use of text, images, and existing 3D CAD models in simulations. This innovative solution goes beyond basic 3D model rendering by embedding essential information, such as physical properties and joint connection structures, directly into 3D assets. As a result, manufacturing environments, robots, components, and workspaces can be optimized for digital twin applications and robot training.

The integration with NVIDIA's Omniverse, which is based on OpenUSD, allows for enhanced 3D workflows, while the Isaac Sim provides advanced robotics simulation capabilities. By leveraging these technologies, manufacturers can create virtual factory environments using their existing data, allowing for the training, validation, and optimization of robotic operations before actual deployment. This streamlined pipeline helps to bridge the 'sim-to-real gap' that often challenges industries during automation transitions.

Expanding Reach in Asia-Pacific and Beyond

As an NVIDIA Inception partner, Endlight showcased its TRINIX-based SimReady 3D pipeline at the NVIDIA GTC event in Taipei earlier this month. The startup is currently collaborating with major manufacturing companies in South Korea to establish digital twins of factories and create simulated environments for robotic learning. Additionally, Endlight plans to expand its supply chain for SimReady assets into the Asia-Pacific manufacturing market, including Taiwan and Japan, as well as the United States.

Endlight's CEO, Jin-young Park, emphasized the importance of building high-quality simulation environments for robots to learn effectively, stating, “For Physical AI to proliferate, it is essential to convert real-world manufacturing assets into SimReady data. We aim to lead the transition to Physical AI in manufacturing by providing a robust data infrastructure connected to NVIDIA's Omniverse and Isaac Sim.”

This partnership marks a significant step for Endlight as it positions itself at the forefront of the evolving Physical AI landscape, promising to enhance the capabilities of manufacturing and robotics through innovative simulation technologies.

Source: Platum. This is an original English briefing based on the cited report.