When Taiwanese media reported on Nvidia CEO Jensen Huang dining with TSMC, supply chain partners, and AI server company executives, curiosity arose about the VIPs accompanying Huang during his visits to three Nvidia branch offices in China and his attendance at their lunar year-end parties
According to Chinese media The Economic Observer, Huang arrived in Beijing on January 19 for Nvidia's branch annual meeting, where he dined with Xingxing Wang, CEO of humanoid robot maker Unitree Robotics, and He Wang, founder of Galbot. Both are representatives of a younger generation of Chinese tech entrepreneurs born in the 1990s, now in their early 30s.
Chinese media ChinaStarMarket also reported that Xingxing Wang shared a photo with Huang on social media, captioned: "New year, new beginning, let's go!" The report highlighted that Huang held meetings with high-level representatives from several leading Chinese robotics companies during his time in Beijing. Aside from Unitree Robotics and Galbot, attendees included executives from LimX Dynamics, Booster Robotics, and Kecheng Huang, co-founder of Emerging AI.
Another media outlet China Entrepreneur noted that Huang and Unitree's Wang are not strangers to each other. In March 2024, during the GTC conference, Huang showcased nine humanoid robots, including those from Unitree. At CES 2025, Nvidia also announced its partnerships with Chinese robotics companies such as Unitree Robotics and XPeng Robotics.
Building comprehensive robotics ecosystem
During his CES 2025 keynote, Huang declared that the "ChatGPT moment for general-purpose robots is coming." Nvidia is actively driving advancements in robotics technology, committed to building a comprehensive robotics ecosystem.
At the event, Huang showcased 14 humanoid robots, six of which were developed by Chinese companies, including Unitree Robotics, AgiBot Innovation Technology, Fourier Intelligence, XPeng Motors, Galbot, and Robotera.
Huang's vision aligns with his late-2024 remarks during a discussion with AI expert Harry Shum at the Hong Kong University of Science and Technology. He emphasized that three types of robots--cars, drones, and humanoid robots--have the greatest potential for mass production, with humanoid robots leading the way.
Deepu Talla, NVIDIA's vice president of robotics, clarified that Nvidia does not directly manufacture robots, and instead it collaborates with robotics manufacturers and solution developers to create a robust platform. This platform integrates three computing systems, software tools, and workflows to support experts in training, testing, and deploying robotic solutions.
One example is Galbot's collaboration with Nvidia. Galbot founder and CTO He Wang detailed how the company leverages Nvidia's Isaac platform and Omniverse technology to generate large volumes of robotic operation data through simulators. This synthetic data is then used to train robot models.
In this process, simulators and rendering engines play a critical role in verifying the physical accuracy of the synthesized scenes. If the scenes are correct, parallel renderers can generate billions of datasets for robot training.
Nvidia's platform systematically addresses the challenges of robot development. Virtual simulation drastically shortens development timelines from an estimated 50 years to just five years, while the open platform provides a comprehensive toolchain for hardware and software, empowering partners and fostering a thriving robotics ecosystem.