26 August 2025 | News
Image Courtesy: Public Domain
Galbot, a global leader in embodied intelligence and robotics, is integrating NVIDIA Jetson AGX Thor into its G1 Premium robot, recently shown at the World Robotics Conference (WRC). As one of the early adopters in the industry, Galbot is leveraging Jetson Thor to elevate performance—unlocking swifter, smoother, and more intelligent autonomy—while setting new benchmarks for deployment across retail, healthcare, and logistics sectors.
Galbot Opens a New Era of Robotics with Jetson Thor
By integrating NVIDIA Jetson Thor, Galbot has significantly raised the bar for general-purpose robotics, accelerating its roadmap and demonstrating how cutting-edge compute translates directly into stronger autonomy and broader commercial deployment.
Built on Jetson Thor, the G1 Premium has already achieved major gains in speed, fluidity, and real-time reasoning. With 7.5x the AI compute of the previous generation NVIDIA Jetson Orin and 3.5x greater energy efficiency, Galbot’s robots can now execute complex planning and motion with a level of precision that redefines what’s possible in embodied intelligence.
“Our G1 Premium, now running on NVIDIA Jetson Thor, has already demonstrated remarkable advancements in speed and improved real-time reasoning capability,” said Professor Wang He, Founder & CTO of Galbot. “This early adoption empowers us to push our proprietary VLA models to new levels of real-world capability.”
The impact of this integration was on full display at the World Robotics Conference (WRC), where Galbot’s G1 Premium showcased fluid motion control, seamless navigation, and precision drifting—earning recognition from attendees as the “swiftest humanoid robot worker.” This demonstration underscored Galbot’s leadership in pushing the limits of embodied intelligence for real-world applications.
Galbot Clinches Gold at the World Humanoid Robot Games
Competing fully autonomously without any reliance on teleoperation, Galbot ranked first in every round, from preliminary to final, ultimately securing the gold medal with a flawless time of 10 minutes and 22 seconds—scoring 336 points, more than 160 points ahead of the runner-up. In the Challenge, Galbot successfully identified and retrieved prescribed medications from multi-tiered shelves with speed and precision, showcasing its unique ability to translate embodied AI into real-world, high-stakes scenarios such as healthcare.
Galbot’s Technological Foundation: Sim2Real and Embodied AI Foundation Models
At the core of Galbot’s achievements is its unique Sim2Real methodology that leverages NVIDIA Isaac Sim and Isaac Lab, open-source robotics simulation and learning frameworks, built on NVIDIA Omniverse. By using massive pre-training on large-scale, high-quality synthetic datasets, refined with only minimal real-world data. This approach drastically reduces dependence on costly real-world collection while significantly improving generalizability across complex environments.
Building on this foundation, Galbot has developed its proprietary VLA technology—GraspVLA, TrackVLA, and GroceryVLA—the world’s first family of end-to-end embodied AI foundation models. Together, they empower Galbot robots with zero-shot grasping, autonomous navigation, and high-density retail operations, enabling fluid, human-like performance in dynamic environments.
By leveraging NVIDIA’s advanced simulation frameworks, Galbot has created the world’s first datasets at the tens-of-billions scale for grasping operations and deformable object manipulation. These extensive synthetic datasets—covering rigid and flexible objects across various physical contexts—ensure that Galbot’s VLA large models generalize effectively and perform with exceptional robustness in real-world applications.
OpenWBT_Isaac: Advancing Humanoid Robot Teleoperation
As a joint effort between Tsinghua University, Shanghai Qi Zhi Institute, and Galbot, OpenWBT_Isaac is introduced as a specialized platform for whole-body teleoperation of humanoid robots in simulation. Built on NVIDIA's powerful Isaac Sim and Isaac Lab simulation capabilities, OpenWBT_Isaac inherits the full-body teleoperation advantages of the OpenWBT system (the world’s first open-source whole-body teleoperation system for humanoids, developed by the same team; GitHub link).
By integrating NVIDIA L20/RTX 5880 Ada GPU acceleration and virtual-real fusion simulation technology, OpenWBT_Isaac enables rapid deployment, high-efficiency data collection, and cross-platform compatibility for multiple robot models. This innovative platform provides robust support for translating research breakthroughs into real-world applications, advancing humanoid robotics capabilities with unparalleled efficiency.
From Technology to Real-World Deployment
Powered by this Sim2Real methodology and large VLA models, Galbot robots are already operating in more than 10 smart pharmacies in Beijing powered by Galbot G1, with plans to expand to over 100 locations nationwide by the end of this year.These robots deliver frictionless, fully autonomous services, showcasing how synthetic data-driven embodied intelligence translates into real-world impact.
In addition, Galbot has partnered with leading manufacturers such as Bosch Group to deploy its embodied AI-powered robot platform in smart production, replacing traditional pre-programmed robotic arms with more intelligent, adaptable solutions.
