NdotLight 入选 Scale-Up TIPS 计划，推进物理 AI 数据基础设施
NdotLight Chosen for Scale-Up TIPS Program to Advance Physical AI Data Infrastructure < IT·Gaming < 기사본문 - The Elec Inc.
NdotLight 于 10 月 13 日宣布入选 2026 年 Scale-Up TIPS 研发计划，将在三年内获得最高 30 亿韩元政府研发资金，用于升级其 3D 数据生成平台 TRINIX。TRINIX 可从文本、图像等输入生成 3D CAD 模型并自动推断质量、摩擦等物理属性，输出可直接用于仿真的 SimReady 资产，以解决遥操作数据采集慢、成本高、质量不稳的问题。其数据客户包括 Holiday Robotics、AROBOT、Robros 和 RLWRLD，并曾为现代汽车、LG 电子的物理 AI 项目供数据。
正文
当前显示原文，中文全文译文尚未就绪。
Company to Receive Up to 3 Billion Won over 3 Years
Hyun-Seon, Park; Published 2026.10.13 17:46; Reply 0
NdotLight, a company specializing in physical artificial intelligence (AI) data infrastructure, said on Oct. 13 that it has been selected for the 2026 Scale-Up TIPS Research and Development (R&D) program, a specialized track run by the Ministry of SMEs and Startups and the Korea Technology and Information Promotion Agency for SMEs.
The company plans to use up to 3 billion won in government R&D funding over three years to enhance TRINIX, its 3D data generation platform.
NdotLight was previously selected for the TIPS program and the Baby Unicorn initiative. Scale-Up TIPS is a government program providing R&D funding to promising technology companies identified and funded by private-sector operators.
Through the project, NdotLight will build a system to compare the physical properties of generated assets with actual measurement data, verify their accuracy, and make corrections. By adding a verification stage to its existing data generation and extraction technologies, the company aims to improve the reliability of simulation data used in robot training and evaluation.
TRINIX is a platform that generates 3D computer-aided design (CAD) models from text, images, and other inputs, while automatically inferring their physical properties. The generated models can be exported as simulation-ready (SimReady) assets for direct use in simulations.
Teleoperation, or remote control, is widely used to collect data for physical AI, but faces limitations such as slow data collection, high costs, and inconsistent data quality. NdotLight said TRINIX addresses these challenges through an automated pipeline for generating 3D data that accurately incorporates physical properties such as mass and friction, as well as joint structures and collision boundaries.
NdotLight's AI training data customers include Holiday Robotics, AROBOT, Robros, and RLWRLD. The company has also supplied data for physical AI projects at major companies including Hyundai Motor and LG Electronics.
NdotLight has raised a cumulative 25.4 billion won, including its recent investment from Naver D2SF. The Series B funding round, led by Korea Development Bank, totaled 15 billion won. Naver D2SF previously participated in the company's pre-Series A round in 2021 and Series A round in 2022. Its latest investment reflects the view that NdotLight can address data bottlenecks that remain a key challenge for physical AI.