ABB and NVIDIA: Physical AI Shapes Intelligent Automation
6 min read
ABB Robotics has announced a breakthrough partnership with NVIDIA that integrates NVIDIA Omniverse libraries directly into ABB Robotics' RobotStudio — the world's most widely used robot simulation and programming software, with more than 60,000 users globally. The integration, unveiled at NVIDIA GTC 2026, delivers industrial-grade Physical AI at scale by closing the long-standing "sim-to-real" gap — the decades-old barrier that has prevented manufacturers from reliably training robots in simulation and deploying them on the factory floor with confidence.
The Sim-to-Real Gap: A Decades-Old Barrier Now Closed
The "sim-to-real" gap refers to the persistent discrepancy between how robots behave in simulation and how they behave in the real world — a problem rooted in the difficulty of accurately replicating real-world lighting, material surfaces, textures, shadows, and physical dynamics in virtual environments. For decades, this gap forced manufacturers into thousands of costly physical test runs, expensive prototyping cycles, and time-consuming debugging processes before new robotic applications could be deployed on production lines.
By integrating NVIDIA Omniverse's photorealistic, physically accurate simulation into RobotStudio, ABB and NVIDIA have now achieved up to 99% correlation between simulated and real-world robot behaviour — a figure that makes simulation-trained deployment commercially viable at industrial scale for the first time. Robots can now see, learn, and understand in simulation, then translate that learning with near-perfect precision to the physical production floor.
"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. For more than 50 years, ABB Robotics has led the evolution of intelligent industrial automation — today's announcement with NVIDIA brings physical AI to industry at scale."
— Marc Segura, President, ABB Robotics
RobotStudio HyperReality: The New Unified Platform
The centrepiece of the partnership is RobotStudio HyperReality — a unified workflow that exports a fully parametrised robot station (robots, sensors, lighting, kinematics, and parts) as a USD file into NVIDIA Omniverse. Inside Omniverse, ABB Robotics' virtual controller runs the exact same firmware as the physical robot hardware — an architectural distinction that is unique to ABB and the foundational reason the 99% sim-to-real accuracy is achievable. No other robot manufacturer currently has a virtual controller running identical firmware to its physical counterpart.
Synthetic images generated within the Omniverse environment feed directly into AI training pipelines, allowing vision models to be trained entirely in simulation — without requiring physical robots or production-line access during the training phase. This combination of physics-rich simulation, synthetic data generation, and ABB's Absolute Accuracy technology — which reduces physical robot positioning errors from 8–15 mm down to approximately 0.5 mm — delivers a system precision level previously unachievable in robotics simulation.
RobotStudio HyperReality is expected to be commercially available in the second half of 2026, fundamentally changing the economics and timelines of industrial automation deployment — reducing production costs by up to 40% and accelerating time-to-market by up to 50%.
"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. Integrating NVIDIA Omniverse libraries into RobotStudio brings advanced simulation and accelerated computing to ABB Robotics' unique virtual-controller technology."
— Deepu Talla, Vice President of Robotics and Edge AI, NVIDIA
Foxconn: The First Industrial Pilot in Consumer Electronics Assembly
The first real-world pilot of RobotStudio HyperReality is being conducted by Foxconn — the world's largest electronics contract manufacturer — in consumer electronics assembly. Consumer electronics assembly is one of the most demanding environments for robotic automation: multiple device variants require different production methods, delicate metal structures demand precise pick-and-place control, and fine-tuned setup typically requires extensive physical debugging cycles.
Through RobotStudio HyperReality, Foxconn's assembly robots are trained entirely in the virtual environment using synthetic data — perfecting multiple real-world production processes across diverse scenarios before a single robot moves to the production line. The result is 99% accuracy at deployment, enabling parallel engineering for better designs, faster production ramp-up, and greater product evolution through advanced AI inference.
"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. We're incredibly excited by the potential of this collaboration, which enables parallel engineering for better designs, faster production ramp-up and greater product evolution."
— Dr. Zhe Shi, Chief Digital Officer, Foxconn
Reaching SMEs: WORKR Extends the Technology to Small Manufacturers
While the Foxconn pilot demonstrates the platform's capability at the highest tier of global manufacturing, California-based WORKR is extending the same technology to small and medium-sized manufacturers across the United States — organisations that have historically been priced out of advanced robotics automation. WORKR's proprietary WorkrCore™ AI platform combines ABB industrial robots trained with NVIDIA Omniverse synthetic data to deploy advanced automation systems that can onboard new parts in minutes and be operated without any programming expertise.
The significance of this reach extension cannot be overstated. Advanced automation has historically been a capital advantage reserved for large manufacturers with the engineering resources and investment capacity to absorb high deployment costs. By democratising access to Physical AI-trained robotics, the ABB–NVIDIA ecosystem is beginning to address one of manufacturing's most structurally important challenges — the critical labour shortages affecting smaller producers who cannot compete for workers at prevailing wages while remaining cost-competitive.
Edge AI and the Road Ahead
The partnership extends beyond simulation into the physical control architecture. ABB Robotics is assessing the integration of the NVIDIA Jetson edge computing platform into its Omnicore robot controller — a move that would enable real-time AI inference at the edge across ABB's entire robot portfolio. Combined with the broader NVIDIA ecosystem momentum — which includes Omniverse-based partnerships with Hyundai, Samsung, Siemens, and Deloitte — the ABB collaboration positions RobotStudio HyperReality as a cornerstone capability for the next generation of AI-native manufacturing infrastructure.
Key Takeaways
- ABB Robotics has integrated NVIDIA Omniverse libraries into RobotStudio, closing the sim-to-real gap with up to 99% simulation-to-deployment accuracy — a first in industrial robotics at scale.
- The new RobotStudio HyperReality platform, available H2 2026 to 60,000+ RobotStudio users, promises to reduce production costs by up to 40% and accelerate time-to-market by up to 50%.
- ABB is the only robot manufacturer running identical firmware in both its virtual controller and physical hardware — the architectural foundation enabling the 99% sim-to-real accuracy claim.
- Foxconn is piloting the first joint use case in consumer electronics assembly; WORKR is extending the technology to SMEs across the US — addressing critical labour shortages without requiring programming expertise.
- NVIDIA Jetson edge computing integration into ABB's Omnicore controller is under assessment — pointing toward real-time AI inference at the edge across ABB's entire global robot portfolio.