StartupsNov 10, 2026, 10:04 AM UTC·2 min read

Ndotlight Supplies KT With Cable Simulation Data for Robots

Korean physical AI data firm Ndotlight is supplying KT with simulation data for cables and other flexible objects, so robots can practice tasks virtually before deployment.

A robotic arm handling a flexible cable in a 3D simulation environment on a monitor in an industrial setting. — Image: 스타트업 생태계의 모든 것 \'플래텀(Platum)\'

Korean physical AI data company Ndotlight says it is supplying KT with 3D simulation data for cables and other deformable objects, letting robots rehearse handling them in a virtual space before they are sent to real industrial sites. The company is expanding this made-to-order robot simulation business on the back of its technical cooperation with Nvidia.

Why flexible objects are hard

For a robot to work in industry, the equipment and surroundings it handles must be reproduced faithfully in a virtual environment, where tasks can be repeated, learned and verified. Flexible objects such as cables change shape when touched or moved, so the data must capture not only appearance but also physical properties such as weight and elasticity, along with their motion. That makes them difficult to build.

Ndotlight runs an on-demand business it calls SimReady assets: 3D data that includes physics information and can be dropped straight into a simulation. For KT, it created cable and other flexible-object data needed at KT's industrial sites. Equipment shapes and part structures are built from CAD drawings, and the bending and movement of flexible objects are modeled so teams can examine how a robot and its work target interact.

TRINIX and Nvidia's ecosystem

The data is produced with TRINIX, a CAD platform Ndotlight developed in-house. It handles 3D model generation, physical property and joint input, collision geometry and export of simulation files in one flow. Output is in OpenUSD, a standard 3D data format, so it can be used in Nvidia's Isaac Sim to build and validate robot task scenarios.

Ndotlight also plans to provide its data to SimReady Central, a SimReady data repository run by Nvidia, making industrial equipment and flexible-object data available to physical AI developers worldwide as part of its overseas push. At NVIDIA AI Day Seoul 2026, held at COEX in Seoul, it presented with Nvidia's headquarters under the theme "From CAD to SimReady," covering the KT case and the SimReady Central plans. It also introduced DexBench, a robot hand-manipulation evaluation project run with Nvidia and Realworld.

CEO Park Jin-young said accurate simulation data of real equipment and work environments is essential for physical AI to spread across industry, and that the company aims to grow into a global physical AI data firm by linking its CAD and AI technology with Nvidia's ecosystem.

Apps worth a look

SUMO: the number puzzle

Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.

Dialo: learn languages by speaking

Talk to an AI conversation partner; every sentence comes back corrected and explained in your language. 9 languages, A1 to C2.

ediviewer: read EDI files like documents

Open EDIFACT and X12 messages, see every segment explained, convert to JSON/XML, export PDF and Excel. Runs locally, no upload.

giftdetect: the right gift in minutes

Pick who it is for, the occasion and their personality. The search engine reads product descriptions from local shops and ranks what fits. Free, no sign-up.

Tags:ndotlightktnvidiaisaac simphysical aisimulation data