During the World Robot Conference 2026, Orbbec and Litchi Lab, a physical AI infrastructure provider, announced a strategic partnership. The two companies will collaborate on world model data collection and training, focusing on the provision of high-quality 3D depth data, as well as the bidirectional alignment and unification of visual sensors and simulated worlds. Together, they aim to advance the implementation of “3D visual perception + physical world model” solutions and improve the quality of training data for world models.
The agreement was signed by Kevin Li, General Manager of Scanning and Innovation Business Unit at Orbbec, and Wei Qin, Chief Marketing Officer at Litchi Lab, and witnessed by Howard Huang, Founder, Chairman and CEO of Orbbec, Yiming Li, Founder of Litchi Lab, and Haozhe Lou, Co-founder of Litchi Lab.
Yiming Li, Founder of Litchi Lab (left), and Howard Huang, Founder, Chairman and CEO of Orbbec (right)
Litchi Lab is committed to building intelligent infrastructure for the Physical AI era. By breaking down the barriers between data, models, and hardware, the company enables the same intelligent capabilities to flow across different robot platforms, supporting reuse, evolution, and expansion into real-world applications including industrial, agricultural, and public service scenarios.
Under this strategic partnership, Orbbec will provide Litchi Lab with chip-level depth camera-based data collection hardware devices required for world model training. Litchi Lab will provide the algorithm foundation for the joint solution and open its self-developed physical world model and data pipeline capabilities to the industry.
The two companies will carry out technical collaboration to promote the bidirectional alignment and unification of visual sensors and simulated worlds, achieving consistency between real-world data and simulation environments at the perception level. Together, they will develop an integrated “3D visual perception + physical world model” solution.
Litchi Lab’s Founder Yiming Li stated that the potential of physical world models to understand physical laws depends on the physical information embedded in training data. Orbbec brings extensive experience in chip-level 3D depth cameras, while Litchi Lab contributes its expertise in physical AI algorithm foundations. Through collaboration, the two companies will enhance world model training quality, enabling AI to not only “see” the physical world, but also “understand” physical laws and ultimately “transform” the physical world. In the future, both parties aim to connect with more robot platforms to serve more industries.
Orbbec’s Founder Howard Huang stated that high-quality, standardized visual hardware is essential for the large-scale collection of multimodal data from the physical world and serves as a foundation for accelerating the deployment of physical AI and embodied AI. Litchi Lab is building an open and unified foundation for world model intelligence, while Orbbec continues to build an open hardware foundation for embodied data collection. The collaborative innovation between the two infrastructure providers will further accelerate the integration and evolution of visual perception and physical world models, empowering more applications including embodied AI scenarios.
