Agibot Showcases Embodied AI Ecosystem and Unveils Simulation Platform at CES 2026
Agibot presented its full lineup of general-purpose robots and introduced a new simulation platform at CES 2026, signaling its intent to accelerate the commercialization and technical iteration of embodied intelligence technologies.
The company’s presence at the Las Vegas Convention Center from January 6-9 highlighted a strategy focused on integrating robotics into diverse industries and daily life. Yao Maoqing, Partner and Senior Vice President at Agibot, stated that the company’s ecosystem is not designed for single tasks but aims to function across various environments.
Full Product Matrix on Display
Agibot exhibited several robot models targeting different operational scenarios. The "Expedition A2" demonstrated multilingual interaction capabilities and performed dance routines, while the "Lingxi X2" showcased stability through Tai Chi movements.
In service applications, the "Genie G2" operated as a waiter, delivering beverages by combining interactive intelligence with high-precision force control. For industrial and outdoor tasks, Agibot displayed the "Kutuo D1" series of quadruped robots. Equipped with reinforcement learning (RL) algorithms, the D1 series is designed for inspection and operations in complex terrains. The D1 Maxpro model demonstrated its payload capacity with audience participation.
The company also introduced the "Qiyuan Q1," a compact, full-body force-controlled humanoid robot capable of being packed into a backpack. Released shortly before CES, the Q1 features "QDD" joints the size of eggs, balancing portability with durability. Additionally, Agibot displayed its "OmniHand" dexterous hand series, emphasizing fine manipulation capabilities essential for complex tasks.
Launch of Genie Sim 3.0
On the opening day of CES, Agibot launched Genie Sim 3.0, an open-source simulation platform driven by large language models (LLMs). Built on NVIDIA Isaac Sim, the platform integrates 3D reconstruction and visual generation to create high-fidelity digital twin environments.
According to the company, Genie Sim 3.0 introduces LLM-driven scene generalization technology capable of generating tens of thousands of scenarios in minutes. Agibot announced it is open-sourcing over 10,000 hours of simulation datasets containing real-machine operation scenarios, alongside a multi-dimensional evaluation system covering more than 100,000 scenarios.
The platform offers a closed-loop solution encompassing digital asset generation, scene generalization, data collection, and automated evaluation. This architecture aims to reduce reliance on physical hardware and accelerate model training and validation for developers. All related assets and source codes have been made available on GitHub.
Partnership and Industry Recognition
During the CES 2026 keynote by NVIDIA Corp., Agibot was the only Chinese robotics company cited by CEO Jensen Huang. Huang acknowledged Agibot’s commercial applications in real-world scenarios and its capabilities in data collection and training.
NVIDIA subsequently confirmed the release of Genie Sim 3.0, which deeply integrates its Isaac Sim technology, in an official press release, underscoring Agibot's position as a key global ecosystem partner.
Agibot executives indicated that the company will continue to focus on technical R&D and ecosystem construction to deepen the practical application of embodied intelligence across various sectors.