Universal Robots demos new physical AI with NVIDIA, Inbolt at GTC Paris

Taking place for the first time, GTC Paris is part of Viva Technology - Europe’s biggest startup and tech event - and UR is presenting cobot demos featuring two of its newest products at the NVIDIA Robotics booth. Namely the all-new UR15, UR’s fastest cobot to date delivering up to 30% cycle time improvements, and the UR AI Accelerator, showcased in Paris for the first time in Europe with commercially viable partner-developed solutions.

The AI Accelerator, UR’s plug-n-play toolkit for developing AI-powered applications, was launched last fall and provides developers with an extensible platform to build applications, accelerate research and reduce time to market of AI products. It is developed in collaboration with NVIDIA using NVIDIA Isaac CUDA-accelerated libraries and models and running on the NVIDIA Jetson AGX Orin system-on-module.

“At Universal Robots, we’re committed to being the leading platform for physical AI, addressing the challenges that were previously unsolvable in close collaboration with our vast ecosystem of partners“, says Anders Billesø Beck, UR’s Vice President, Technology. “The demos we’re presenting at GTC Paris showcase just that. And the best part is that they’re not prototypes or future visions. It’s tested, off-the-shelf technology ready to be deployed and provide value to manufacturers today.”

On June 12 at 13:00, Anders Billesø Beck will also be part of a GTC Paris panel discussion with other robotics leaders, discussing how physical AI is shaping the next generation of industrial robots.




Universal Robots demos at GTC Paris
Debuting at a major event in Europe, a UR15 running on UR’s PolyScope X software is used to demonstrate how applications developed with the AI Accelerator can easily handle variations within unpredictable and changing environments. The pick-and-place application knows what objects to pick but not their position and with input from the mounted camera and the AI Accelerators incorporated NVIDIA AI technology, the cobot localizes and grabs the object, then placing it in a random position to repeat the process. This is especially valuable for e.g. machine tending and bin picking applications in factory environments or in logistic centers where variations naturally occur.

On a UR5e cobot and fully integrated in the UR AI Accelerator toolkit, Inbolt is showcasing the only vision-guidance solution capable of enabling intelligent robot decision making in dynamic and unstructured environments. Inbolt technology is deployed by global manufacturers like Ford, Toyota, Stellantis for pick & place, screw driving, dispensing applications and more.

The high-speed, robot-mounted vision-guidance solution continuously tracks and adjusts the robot’s trajectory based on part position and orientation, even in dynamic or low-visibility environments.




The NVIDIA booth is found at Hall 7.3.
Elsewhere at GTC Paris, Universal Robots is teaming up with Accenture and NVIDIA to demonstrate the next generation human-machine workforce in life science manufacturing. In this part digital, part physical demo, cobots become seamless extensions of human operators, performing precise, repetitive, and potentially hazardous tasks. The demo shows advanced reinforcement learning techniques for training and robotic manipulation; how a trained model can be transferred into a digital twin for validation and testing prior to implementation; and finally, the transition to the physical world, where a robot executes tasks in response to voice commands from a human operator.

The Accenture booth is found at P05.

www.universal-robots.com