ABB Robotics and Nvidia on Monday revealed a new partnership that they say will be instrumental in closing the sim-to-real gap in industrial robotics.
Under the deal, the companies will combine Nvidia’s Omniverse simulation capabilities with ABB’s software design and simulation suite, RobotStudio, to help manufacturers deploy physical AI in real-world robotics applications.
The resulting product, dubbed RobotStudio HyperReality, will be available in the second half of this year under a subscription model. The vendors claim it can cut engineering time in half, reduce deployment costs by 40%, and accelerate time-to-market by up to 50%. An initial pilot of the platform is currently underway by Foxconn for consumer electronics assembly, as well as by the California-based robotic workforce company Workr.
Closing the Sim-to-Real Gap
The core promise of the platform is a hyper-realistic virtual environment where developers can accurately design and test entire automation processes before deploying them in the real world.
In particular, vendors are pitching the tool as offering training data across previously complex conditions, such as different lighting, textures and material properties. In this way, it claims to offer an effective means of closing the gap between simulation accuracy and real-world conditions, also known as the sim-to-real gap.
According to the partners, the new platform has achieved a 99% correlation between simulation and real‑world behavior.
The simulations are continuously refined with real-world data feedback, meaning the underlying AI models improve over time without requiring thousands of physical test runs or expensive physical prototypes.
Speaking at a media briefing, Marc Segura, President of the ABB robotics division, framed the new platform as a crucial inflection point for the uptake of physical AI.
“For decades, one final barrier has limited the deployment of physical AI at speed and scale,” he said. “Manufacturers have struggled to simulate real-world conditions; lighting, shadows, textures, and materials that behave just like they do on the factory floor. That is known as the sim-to-real gap, and that has limited how quickly businesses can design, test, and deploy new automation. Today, we have broken that barrier.”
Segura described the platform’s benefits as twofold.
Beyond enhancing simulation capabilities, the platform also enables the training of higher-quality AI models that translate directly into greater robot autonomy and versatility. “You’ve got the simulation benefits, and you’ve got the benefit of more autonomy for the robots on one and the same platform,” he added.
Also speaking at the briefing, Deepu Talla, Nvidia’s vice president of robotics and edge AI, said the partnership reflects a broader shift in what robotics can be.
“Physical AI represents the next frontier of intelligence, moving beyond digital models that simply process text or images to embodied models that can perceive, reason, and act within the physical world,” he said, adding that this evolution is transforming robots from fixed, pre-programmed machines into adaptable, autonomous agents capable of operating in complex and dynamic environments.
The partners also highlighted uptake among small and medium-sized businesses as a key aim, as these businesses are typically excluded and underrepresented in robotics rollouts due to financial and technological barriers.
With RobotStudio HyperReality, the partners say these companies will now have access to a user-friendly, cheaper means of testing and using industrial robots.
“The goal with this new product is to take a huge step in making robots available for SMBs,” Segura said.
