Physical AI data company Ndotlight said on Tuesday it raised 15 billion won in funding.

The investment was led by KDB Industrial Bank, with existing investors Naver D2SF, IMM Investment and Capstone Partners taking part in a follow-on round. NH Venture Investment, OurCrowd and Stonebridge joined as new investors.

Ndotlight provides training data for Physical AI based on its in-house simulation 3D asset generation solution, TRINIX.

TRINIX is a technology that automatically generates 3D data reflecting physical properties such as mass and friction, joint structures and collision ranges. It can generate synthetic data at scale for robot training by linking to Nvidia's simulation platform Omniverse.

The company explained that this helps address cost and speed issues in data collection methods such as teleoperation and motion capture in real environments, and reduces the Sim-to-Real Gap, the difference between simulation and reality.

It is currently participating as a data supplier in Physical AI projects at Hyundai Motor and LG Electronics. It also supplies AI training data to robot and robot foundation model companies including Holiday Robotics, Arobot and RealWorld.

Ndotlight said it plans to use the proceeds to upgrade a generative CAD engine and a physical-property database, and to expand integration with Nvidia Isaac Sim and Omniverse workflows. It will also expand its push into global markets and hiring across all job categories.

Ndotlight CEO Jin-young Park (박진영) said, "The key in the era of Physical AI depends on how efficiently high-quality data that can be learned immediately in a simulation environment can be secured." He added, "We will leap into a Physical AI data infrastructure company trusted at global industrial sites."