3D AI company NdotLight has unveiled its technology for generating 3D CAD-based Sim-Ready assets and next-generation Physical AI pipelines on the global stage. NdotLight presented the technology at NVIDIA Startup Day, a startup event held as part of NVIDIA GTC Taipei 2026 in Taipei, Taiwan, on June 2.

The presentation came immediately after NVIDIA CEO Jensen Huang emphasized the importance of the Physical AI ecosystem during his keynote the previous day. NdotLight was also added to NVIDIA’s partner ecosystem list, joining as a technology company addressing one of the key bottlenecks in implementing Physical AI: the generation and simulation of Sim-Ready assets.

Sim-Ready assets are 3D assets designed for simulation that contain physical information, enabling robots and AI systems to learn and validate tasks. NdotLight is automating the process of generating Sim-Ready assets and connecting them to simulation environments through its proprietary AI solution, TRINIX, built on its own 3D CAD engine.

At the event, NdotLight demonstrated technology that generates 3D CAD-based Sim-Ready assets from text, images, and existing CAD data and integrates them with major simulation environments such as NVIDIA Isaac Sim. Going beyond the generation of complex articulated objects, which has been a key focus of its technology, NdotLight also unveiled technology for generating and simulating deformable assets whose shapes physically change.

During the on-site demonstration, the company showcased simulations of the connection and fastening processes of power cables and LAN cables, which are considered challenging areas in manufacturing and automation processes, as well as the physical deformation of rigid bodies and soft bodies. The technology can reduce the time required to build complex physical-environment assets needed for robot and Physical AI training, drawing interest from industry professionals in manufacturing, robotics, and digital twins.

Park Jin-young, CEO of NdotLight, said, “For the era of Physical AI to fully take off, it is essential to build simulation environments in which robots and AI can learn quickly and precisely. NdotLight is building infrastructure that converts 3D data into Sim-Ready assets using our proprietary 3D CAD engine and AI technologies and connects them so they can be utilized in real-world simulation environments.”

She added, “By strengthening our integration with global simulation ecosystems, including NVIDIA’s, we will contribute to enabling companies in manufacturing, robotics, and digital twins to build Physical AI environments more quickly.”