ABB Robotics has announced the integration of Nvidia Omniverse libraries into its simulation and programming platform, Robot Studio, with the aim of significantly accelerating the scalability of industrial production. This integration paves the way for widespread industrial adoption of physical AI. By combining ABB’s robotic software with highly accurate simulation, the previously existing “Sim-to-Real” gap between virtual development and real production is nearly closed. By linking digital twins with realistic simulations, robots can achieve up to 99 percent accuracy between the model and its application, as stated in a press release.
Developers can design robot processes entirely in a virtual environment, generate synthetic training data, and train AI models prior to their deployment. ABB leverages its Virtual Controller, which precisely mirrors real control software, along with the Absolute-Accuracy technology that offers positioning accuracies down to 0.5 millimeters. A key element of this collaboration is Robot Studio Hyper Reality, expected to be available from the second half of 2026. This platform aims to reduce setup times by up to 80 percent, cut costs by up to 40 percent, and accelerate the introduction of new products by up to 50 percent.
Initial pilot applications are already underway, including at Foxconn in the assembly of consumer electronics, where robot processes are initially optimized entirely virtually. ABB and Nvidia are also exploring the integration of Edge AI through the Jetson platform into ABB controllers to execute AI functions directly on the robot. At Nvidia GTC, robotics company WORKR is showcasing applications that utilize ABB robots trained on synthetic data to enable automation without programming effort, particularly benefiting small and medium-sized manufacturers and addressing labor shortages.
About the Author
