ABB Robotics and Nvidia just solved one of industrial automation's most stubborn problems: training robots in simulation so accurately that they work flawlessly in the real world. The partnership integrates Nvidia's Omniverse libraries into ABB's RobotStudio software, creating hyper-realistic digital simulations where robots can learn complex manufacturing tasks before ever touching a production line. This breakthrough addresses what engineers call the "sim-to-real gap"—the long-standing accuracy deficit between virtual training environments and actual factory conditions. What Is the Sim-to-Real Gap and Why Does It Matter? For decades, manufacturers have struggled with a fundamental problem: robots trained in computer simulations often fail or perform poorly when deployed in real factories. The gap exists because virtual environments can't perfectly replicate real-world lighting, material textures, surface variations, and environmental conditions. This forces companies to spend months physically testing and retraining robots on actual production lines—a costly and time-consuming process. "The industrial sector needs physically accurate simulation to bridge the gap between virtual training and the real-world deployment of AI-driven robotics at scale," explained Deepu Talla, vice president of robotics and edge AI at Nvidia. The new RobotStudio HyperReality system delivers unprecedented accuracy by combining Nvidia's physically accurate simulation power with ABB's unique virtual controller technology—the only system that runs the same firmware in simulation as it does in actual hardware. How RobotStudio HyperReality Works and What It Delivers The collaboration creates a game-changing workflow for manufacturers. Developers can now simulate robots in digital twins and generate synthetic training data using Nvidia Omniverse libraries. These foundation models are continuously optimized with real-world feedback, improving system performance over time. ABB's Absolute Accuracy technology reduces positioning errors from 8–15 millimeters down to around 0.5 millimeters, ensuring precision that meets industrial-grade requirements. The results speak for themselves. By designing, testing, and optimizing production lines virtually, manufacturers can achieve measurable efficiency gains: - Setup and Commissioning Time: Reduced by up to 80% by eliminating the need for extensive physical testing and debugging - Cost Reduction: Cut by up to 40% through elimination of physical prototypes and reduced engineering labor - Time-to-Market: Accelerated by 50% for complex products like consumer electronics - Simulation Accuracy: Achieves up to 99% accuracy, enabling robots to transfer learning directly to production floors Real-World Pilots Show Immediate Impact The technology isn't theoretical—major manufacturers are already testing it. Electronics contract manufacturer Foxconn is piloting the first joint use case in consumer electronics assembly, where precision is critical. Assembling tiny components in consumer electronics requires handling multiple device variants with different production methods, precise pick-and-place control, and fine-tuned setup. Using RobotStudio HyperReality, Foxconn's assembly robots are trained virtually using synthetic data to perfect multiple real-world production processes before moving to the actual production line with 99% accuracy. "Precision is everything in consumer electronics manufacturing and until now, this level of accuracy and fidelity just wasn't possible in simulation and digital twins," said Dr. Zhe Shi, chief digital officer of Foxconn. "We're incredibly excited by the potential of ABB Robotics and Nvidia's collaboration, which enables parallel engineering for better designs, faster production ramp-up and greater product evolution through advanced AI inference and understanding." California-based robotic workforce company WORKR is extending this technology to small and medium manufacturers across the United States. By combining ABB's industrial-grade robotics with its proprietary WorkrCore AI platform, WORKR is helping manufacturers address labor shortages with robotic systems that can learn new tasks in minutes and be operated by anyone—no programming knowledge required. Steps to Deploy RobotStudio HyperReality in Your Manufacturing Process - Assess Your Current Workflow: Identify production processes where setup time, cost, or precision are major bottlenecks—consumer electronics assembly, automotive manufacturing, and precision machining are ideal candidates - Design Virtual Simulations: Work with ABB to create digital twins of your production lines using RobotStudio, leveraging Nvidia Omniverse libraries for physically accurate environmental modeling - Generate and Train with Synthetic Data: Use the platform to create synthetic training data that covers multiple production scenarios, variants, and edge cases before deploying robots to the physical line - Validate with Real-World Feedback: Deploy trained robots to production and continuously feed performance data back into the system to optimize foundation models over time - Scale Across Your Operations: Once validated, replicate the trained models across your robot fleet worldwide with confidence in 99% accuracy transfer What This Means for the Broader Robotics Industry This partnership signals a major inflection point for industrial robotics. ABB said it will release RobotStudio HyperReality to its 60,000 existing RobotStudio customers worldwide in the second half of 2026, democratizing access to industrial-grade physical AI. The collaboration also builds on ABB and Nvidia's long-standing work together, including previous integration of Nvidia Jetson into ABB's autonomous mobile robots and development of gigawatt-scale AI data centers. Marc Segura, President of ABB Robotics, emphasized the significance: "Today, using Nvidia accelerated computing and simulation technologies, we have removed the last barriers to making industrial and physical AI a reality at a global scale by closing the sim-to-real gap." This breakthrough addresses a problem that has constrained robotics adoption for over 50 years, finally enabling manufacturers of all sizes to deploy AI-driven automation with confidence.
