Schaeffler Transforms Production With Digital Twins and Physical AI on NVIDIA Omniverse

Overview

Schaeffler Group built a digital twin platform on NVIDIA Omniverse and OpenUSD, in collaboration with Wandelbots, to plan, simulate and optimize plants, machines and workflows across its 100+ global manufacturing sites. Pilot projects in e-motor assembly (China), an inverter assembly line (Germany), and Vehicle Lifetime Solutions warehouses (Europe) cut robotic task development time from hundreds of hours to half a day and reduced commissioning times for new production lines; Schaeffler plans to integrate digital twins into more than half its manufacturing sites by 2030.

The challenge

Schaeffler faces rising labor costs and workforce shortages, increased production complexity that demands more sophisticated assembly and quality control, supply chain volatility, and intensifying pressures related to sustainability and regulatory compliance.

The solution

In collaboration with NVIDIA and Wandelbots, Schaeffler built a digital twin platform on NVIDIA Omniverse, using OpenUSD to integrate planning and production data from its industrial expert software, enabling physics-based digital twins for planning, simulating and optimizing plants, machines and workflows across its 100+ global manufacturing sites. The AI computing infrastructure is currently based on a mixed architecture of local virtual machines and NVIDIA L40S GPUs. Engineers use the digital twins to virtually design and optimize facility layouts, simulate material flows and human-robot collaboration, run 3D-based simulations of automated guided vehicles (AGVs), and stream live process data such as energy consumption for monitoring. For robot training, the platform supports simulation-first development of robot fleets — manipulators, autonomous mobile robots and humanoid robots — generating synthetic data to train AI models and reduce the simulation-to-real gap; Wandelbots deployed trained models onto physical robots using a virtual controller for sim-to-real alignment, with reinforcement learning refining robotic movements to within 5-6 centimeters for gearbox assembly tasks.

Reported business value

Schaeffler cut robotic task development time from hundreds of hours to half a day and achieved consistent robotics performance after just 24 hours of training. The platform reduced commissioning times for new production lines and accelerated facility layout planning. Schaeffler plans to integrate digital twins into more than half of its global manufacturing sites by 2030.