MetAI Integrates NVIDIA Omniverse Tech to Accelerate Digital Twin Creation for Industrial AI

Image Courtesy: Public Domain

MetAI,a Taiwan-based member of NVIDIA's Inception startup program, is building AI-powered simulation environments powering the next generation of industrial and physical AI. As part of its mission to accelerate simulation-driven AI development, MetAI is integrating NVIDIA Omniverse technologies, including NVIDIA USD Search microservices, into its digital twin generation platform, MetGen—enabling the development of 3D SimReady environments from 2D CAD blueprints, beginning with the warehousing industry.

MetGen, MetAI's SimReady environment generation engine, is capable of transforming standardized 2D CAD files and structured data into high-fidelity, operational digital twins—ready for AI training, robotics testing, automation logic refinement, and synthetic data generation. With the integration of NVIDIA USD Search, users can now instantly locate and retrieve standardized SimReady assets from connected libraries, significantly reducing build time and enhancing scalability.

"From layout blueprints to SimReady assets, we're building the infrastructure needed to train the next generation of industrial and physical AI," said Daniel Yu, Co-founder and CEO of MetAI. "By integrating NVIDIA USD Search, we're enabling a closed-loop workflow—where users can generate assets with MetGen, build their own searchable libraries, and dynamically retrieve components to streamline simulation creation. This shortens setup cycles and unlocks scalable, intelligent environment generation."

The integration has been validated and is currently in pilot deployment. Key capabilities include:

• Seamless conversion of 2D CAD layout blocks into searchable metadata
• Automatic retrieval of matching 3D assets from connected USD libraries
• Generation of missing assets via MetGen's internal pipeline
• Future integration with agentic AI tools for intelligent and interactive scene construction

This milestone marks a significant step toward realizing Real-to-Sim and Sim-to-Real workflows—starting with warehouse automation and scaling into other industrial domains.

MetAI is already exploring additional applications of this workflow across sectors such as data center construction, advanced manufacturing lines, and robotics training environments. As MetGen prepares for its beta release in August, the MetAI team remains focused on building an extensible ecosystem of SimReady digital twins tailored for industrial AI development.