Industrial AI data infrastructure company NdotLight will participate in Cisco Connect 2026 Korea, taking place on April 8, where it will unveil its 3D data pipeline technology for Physical AI training.

Cisco Connect is Cisco’s flagship annual event for sharing its AI, security, and networking technologies. This year’s key topics include AI infrastructure strategies, digital transformation, and the future of work.

Based on its vision of “Making everything sim-ready,” NdotLight will showcase technology that automatically converts existing industrial 3D data into data suitable for AI training.

At the core of the technology is a pipeline that transforms legacy 3D CAD data into Sim-Ready assets that can be immediately used in robot simulation and digital twin environments. This enables existing data used in industrial settings to be leveraged for AI training without requiring separate processing.

The technology primarily aims to address the “Sim-to-Real Gap,” the discrepancy between simulation and real-world environments, which has been identified as a challenge in manufacturing.

As it is difficult to secure the large-scale datasets required for robot training, NdotLight explains that its technology can efficiently generate synthetic data from 3D data, thereby accelerating the training process. This is considered a key infrastructure component for the expansion of Physical AI and robotic automation.

At this year’s event, the development of next-generation infrastructure in preparation for the expansion of agentic AI and Physical AI will be a major topic of discussion. AI capable of operating in real-world physical environments, rather than simply functioning as software, is emerging as a key factor in industrial competitiveness.

NdotLight’s technology will be presented as a solution addressing this trend by enabling access to high-quality 3D data.

However, practical challenges remain in industrial environments, including data quality, standardization, and the cost of real-world implementation. In particular, how to address variations in the quality of existing CAD data and differences between environments is expected to be a key factor in the broader adoption of the technology.

The extent to which simulation results correspond to real-world environments is also expected to play an important role in companies’ decisions to adopt the technology.

Park Jin-young, CEO of NdotLight, said, “We will support companies preparing to adopt Physical AI so that they can build simulation environments without the burden of excessive time and costs.”

Yoo In-chun, Reporter
ceo@startupn.kr

Source: StartupN
StartupN