Nvidia, Disney, and Google DeepMind's collaboratively developed robot, Olaf, is set to debut in theme parks.

NVIDIA CEO Jensen Huang unexpectedly introduced a special guest at the end of yesterday’s GTC 2026 keynote speech: Olaf, a robot developed in collaboration with Disney and Google DeepMind, making its debut with a surprising appearance. This robot, combining advanced artificial intelligence and physical simulation, demonstrated agile movements and lively gestures. Disney recently announced that Olaf will be featured at Disneyland Paris and Hong Kong Disneyland starting at the end of March.

NVIDIA Teams Up with Google DeepMind to Develop Newton, Teaching Olaf to Walk

The technology enabling Olaf to walk combines hardware computing power with software simulation. The project uses the Newton Physics Engine, jointly developed by Disney Research, NVIDIA, and Google DeepMind. The Newton Physics Engine allows high-performance robots to perform rapid simulations in GPU environments. Through NVIDIA’s GPU-driven simulation programs, the robot can operate stably in real-world settings. To achieve the visual quality seen in animations, Olaf’s body is embedded with shimmering fibers to mimic snow’s gloss, and features a magnetic design that allows its stick-like arms, carrot nose, and hair to be detachable and reassembled. This enables the robot to replicate iconic movements and facial expressions from the film.

In terms of movement design, Disney animators provided extensive AI training data to help the robot learn walking in virtual environments. The development team emphasized not only mobility but also the precise recreation of Olaf’s distinctive “staggering gait” from the animation. This technology overcomes previous limitations where robots’ movements were stiff or lacked character, allowing Olaf to roam freely within the park with a posture closer to the original character. During interactive demonstrations, Olaf displayed smooth dynamic performance, with mouth and eye movements significantly enhancing the realism of interactions.

Operator-Assisted Speech

According to a media preview report by CNET from Disney’s Imagineering headquarters in Los Angeles, Olaf currently requires human assistance to speak. On-site operators can select voice responses based on the situation. While the current speech capabilities are still limited, the robot can perform scripted dialogues. Disney Imagineering engineer Josh Gorin stated that the team has spent years working to bring virtual characters into physical form, and now AI and hardware technologies have made this possible. The robot is not yet ready for physical hugs with visitors, but the team plans to expand interactive features and improve immersive experiences in the future.

Since Olaf’s debut yesterday, it will begin a global tour. The first stop is scheduled for March 29 at the opening ceremony of the “Frozen” themed area at Disneyland Paris. It will also appear at Hong Kong Disneyland, where AI-driven Olaf will interact with visitors. In the future, Disney’s IPs may be transformed into robots that interact with humans, which is great news for Disney fans.

This article about NVIDIA, Disney, and Google DeepMind’s collaborative Olaf robot debuting at theme parks first appeared on Chain News ABMedia.