MetGen SimReady Digital Twin Platform

MetAI is a Taipei-based startup founded in 2023 by Renton Hsu, Daniel Yu and Dave Liu. Backed by NVIDIA and part of the NVIDIA Inception program, it builds a simulation data layer between physical systems and AI models, turning real layouts, behavior and control logic into data for training, validation and decisions in warehouses, semiconductor fabs and data centers.

MetGen, built natively on NVIDIA Omniverse and OpenUSD, converts CAD files and 2D layouts into simulation-ready 3D digital twins in minutes through a CAD-to-Capable pipeline. Parametric SimReady assets carry geometry, materials, physics and logic and can be reused, scenarios can be randomized to cover rare and failure cases, and control systems such as PLCs, ROS, WMS and Python can be connected for testing and AI training.

• CAD-to-Capable pipeline: turns CAD files and 2D layouts into 3D simulation-ready twins.
• SimReady assets: modular, parametric geometry, materials, physics and logic that can be reused.
• Scenario randomization: varies layouts, parameters and operating conditions, including edge cases.
• Control integration: connects PLCs, ROS, WMS and Python for real-time control testing.
• AI training data: generates synthetic data grounded in real layouts and constraints.
• Industry editions: MetGen for Warehouse and MetGen for Semiconductor on NVIDIA Omniverse.

A logistics operator in Vietnam could build a twin of a new automated warehouse from its CAD drawings to test AMR traffic. A semiconductor back-end plant could simulate material flow before adding automation. A system integrator could validate PLC and WMS logic virtually before commissioning.

Digital twins are generated from existing drawings in minutes instead of months of manual 3D modeling. Automation and robotics systems can be validated and iterated before real deployment. Randomized and edge-case scenarios provide training data that would be risky or impractical to collect on a live site, and assets are reused across suppliers and configurations.

Physical AI and robotics need large volumes of realistic simulation data, and Omniverse-based digital twins are becoming the standard way to build it. Smart warehouses and semiconductor capacity expanding in Southeast Asia create demand for faster twin creation before automation is installed.

MetAI was founded in 2023 and is headquartered in Nangang District, Taipei. Its 4 million USD seed round was backed by NVIDIA, Kenmec, Solomon, SparkLabs and Addin Venture. It was named to The Information's 50 Most Promising Startups of 2025, and Digitimes reports it works with TSMC and exhibited at CES 2026.