CtrlK
| Developer | Lanxin Robotics (Hangzhou Lanxin Technology Co., Ltd.) |
| Type | Humanoid robot (research platform) |
| Also known as | VB1-I |
| Country of origin | China |
| Height | 1.65 m (5 ft 5 in) |
| Dimensions | 700 x 610 x 1,650 mm |
| Weight | 65 kg (143 lb) |
| Degrees of freedom | 42 (total); 6 per arm; 7 per hand |
| Arm payload | 2 kg (per hand) |
| Max speed | 1.5 m/s (5.4 km/h, 3.4 mph) |
| Battery life | 2 to 3 hours |
| Navigation | 3D Laser SLAM |
| Driving mode | Dual-wheel differential |
| Price | Approximately $120,000 USD |
| Status | In production |
| Website | lanxinrobotics.com |
The Lanxin VB1-I is a humanoid robot developed by Lanxin Robotics (formally Hangzhou Lanxin Technology Co., Ltd.), a Chinese robotics company headquartered in Hangzhou, Zhejiang Province. Positioned as a dedicated research and development platform, the VB1-I is designed for embodied AI experimentation, logistics trials, and factory simulations. It occupies a middle position in Lanxin's humanoid robot lineup, sitting between the original VersaBot VB-1 and the more advanced VB2.
The VB1-I builds on the vision-first philosophy that Lanxin established with its VersaBot VB-1, the robot the company introduced at the 2024 World Robot Conference as the world's first "pure vision" humanoid robot. While the original VB-1 relied exclusively on RGB-D cameras and rejected LiDAR entirely, the VB1-I adopts a hybrid sensing approach that combines panoramic RGB-D depth vision with 3D Laser SLAM navigation for improved localization accuracy. This shift reflects Lanxin's pragmatic engineering philosophy: deploying the most effective sensor combination for each product's target environment rather than adhering rigidly to a single sensing paradigm.[1][2]
The VB1-I is priced at approximately $120,000 USD and targets research laboratories, AI development teams, and industrial automation integrators who need a human-scale robotic platform for prototyping and experimentation.[3]
Lanxin Robotics was founded in 2016 by Dr. Gao Yong, who holds a doctorate in pattern recognition and artificial intelligence from the Chinese Academy of Sciences. Prior to founding the company, Dr. Gao worked at Ricoh Japan and the Microsoft Asia Engineering Institute on computer vision products. The company's core founding team includes researchers from the Chinese Academy of Sciences, Fudan University, and Zhejiang University, with approximately half of all staff dedicated to research and development.[4][5]
Headquartered at China Artificial Intelligence Town in Hangzhou, Lanxin was the first company in China to develop 3D visual perception technology specifically for mobile robots and to achieve large-scale commercial deployment of that technology. The company operates under an organizational framework it describes as "one body and two wings": the main body is mobile robot systems (MRS), while the two wings are the MRDVS sensor division (3D visual sensor development) and the VMR robot division (visual perception mobile robots, including humanoid robots).[6][7]
By 2024, Lanxin had accumulated 156 national patents, including 41 invention patents, and reported a compound annual revenue growth rate exceeding 80% from 2019 to 2024. The company was listed among Hangzhou's quasi-unicorn companies in 2022 and was recognized as a national-level specialized and new "little giant" enterprise by the Chinese government in 2024.[4][8]
Notable clients for Lanxin's mobile robot and logistics solutions include Huawei, ZTE, BYD, Midea, Toshiba, Foxconn, Toyota, LG, Sharp, Mitsubishi, and COMAC, with deployments spanning the 3C electronics, semiconductor, photovoltaic, battery, automotive, and packaging industries.[1][9]
Lanxin Robotics has completed multiple rounds of financing since its founding. According to Crunchbase, the company raised a total of $15.83 million in disclosed funding as of its Series C round. Key investors include Tencent, Kunpeng Capital, Lanchi Ventures, Advantech Capital, Envision Capital, and Blue Horizon Capital.[10][11]
On May 7, 2025, Lanxin officially announced the completion of its C+ round of financing, led by Kunpeng Fund, with the total amount reported to be in the hundreds of millions of yuan. The company stated that the new capital would fund three priorities: deepening the research and development of 3D vision sensors, advancing the AI technology behind humanoid robots, and upgrading the global supply chain and service system. Lanxin operates two intelligent manufacturing bases, one in Huzhou, Zhejiang Province, and another in Gui'an New District, Guizhou Province.[6][10]
Lanxin has placed an initial public offering (IPO) on its agenda and is expected to pursue listing as what it has described as the "first stock of 3D visual perception robots" in China.[6]
The VB1-I emerged from Lanxin's strategy of iterating on humanoid robot platforms in response to deployment experience and customer feedback. The original VersaBot VB-1 was unveiled at the 2024 World Robot Conference in Beijing on August 23, 2024, where it was introduced as a wheeled-base humanoid with a pure vision navigation system and no LiDAR sensors. The VB-1 was designed for manufacturing logistics, material handling, and smart factory applications, relying on a humanoid upper body with gripper hands mounted on a wheeled mobile base.[1][2]
The VB1-I represents a refinement of this concept, repositioned specifically for research and development use cases. While the VB-1 was optimized for immediate industrial deployment, the VB1-I was designed with modularity and experimentation in mind, allowing researchers to test different control algorithms, manipulation strategies, and AI models on a human-scale platform.[3][12]
The VB1-I stands 1,650 mm (approximately 5 feet 5 inches) tall with overall dimensions of 700 x 610 x 1,650 mm. The robot weighs 65 kg (143 lb) and features a ground clearance of 30 mm and a rotation diameter of 580 mm. Its frame is constructed from aluminum alloy combined with composite materials, balancing structural rigidity with weight reduction.[3][12]
The robot features a humanoid upper body with a head, torso, and dual articulated arms. The upper body provides the human-scale form factor needed for research into human-robot interaction, object manipulation in environments designed for people, and whole-body coordination studies. The operating height range spans 0.2 to 2 meters, giving the VB1-I flexibility to reach both floor-level and elevated surfaces.[12]
The VB1-I uses a dual-wheel differential driving mode for locomotion, as specified on Lanxin's official product page. This differentiates it from the VB-1 (which also uses a wheeled base but with different chassis geometry) and the VB2 (which uses a four-wheel omnidirectional driving system). The dual-wheel differential configuration provides reliable, cost-effective mobility suited to structured indoor environments such as laboratories, test facilities, and factory floors. The maximum travel speed is 1.5 m/s (5.4 km/h).[12]
Some third-party robotics databases have described the VB1-I as featuring bipedal locomotion. However, the manufacturer's official specifications list a dual-wheel differential driving mode, consistent with Lanxin's broader product strategy of using wheeled bases to prioritize stability and reliability in controlled environments.[3][12]
A notable departure from the original VersaBot VB-1 is the VB1-I's adoption of 3D Laser SLAM (Simultaneous Localization and Mapping) for its primary navigation system. The VB-1 was marketed as the world's first "pure vision" humanoid robot, using only RGB-D cameras and no LiDAR. The VB1-I instead uses 3D Laser SLAM for precise localization and mapping, acknowledging that laser-based methods can offer superior accuracy and repeatability in structured research environments where these qualities matter more than the cost savings of vision-only navigation.[12]
The robot also integrates a Panoramic RGB-D Depth Vision System, which provides rich three-dimensional environmental perception for obstacle detection, object recognition, and scene understanding. This combination of laser-based SLAM for navigation with vision-based sensing for perception creates a multi-modal sensor architecture that gives researchers flexibility in how they approach different experimental scenarios.[12]
Lanxin's proprietary LX-MRDVS (Lanxin Mobile Robot Deep Vision System) technology underpins much of the VB1-I's visual processing capability. The MRDVS sensor lineup is organized into four product series:[4][13]
| Series | Function | Key Specification |
|---|---|---|
| S-Series | Obstacle detection (dToF cameras) | 45 m long-range mapping and SLAM |
| M-Series | High-precision docking (iToF cameras) | 0.3 to 5 m range; wide FOV (H-108, V-82 degrees) |
| V-Series | Visual SLAM navigation | Proprietary top-view technology |
| V2 Pro | Integrated spatial intelligence (LiDAR + RGB + IMU) | 3 cm precision real-time localization for AGVs |
The VB1-I is equipped with two articulated arms, each featuring 6 joints (degrees of freedom). The arms can carry a gripping load of 2 kg per hand, which is suitable for handling lightweight objects commonly encountered in research scenarios such as grasping household items, sorting components, and performing pick-and-place tasks.[12]
The total system provides 42 degrees of freedom, distributed across the arms, hands, torso, head, and mobility platform. Each hand has 7 degrees of freedom and features five-finger dexterous hand designs. The five-finger configuration allows for more naturalistic grasping patterns compared to the gripper-style hands used on the original VB-1, supporting research into dexterous manipulation, grasping algorithms, and human-like object interaction.[3]
The robot uses high-torque electric servo motors paired with harmonic drive reducers (also called strain wave gears) throughout its actuator system. Harmonic drives are widely used in precision robotics because they offer high gear ratios in compact form factors, near-zero backlash, and smooth torque transmission. These properties are particularly important in a research platform where repeatable, precise motion is valued over raw payload capacity.[3]
The VB1-I runs on a Linux-based operating system, providing researchers with the flexibility to deploy custom software stacks, ROS (Robot Operating System) packages, and machine learning frameworks. Connectivity options include Ethernet and WiFi, allowing the robot to be integrated into laboratory networks, cloud computing environments, and multi-robot coordination setups.[3]
Lanxin's self-developed core controller serves as the central "brain" of the VB1-I. This controller has been iteratively optimized using operational data collected from Lanxin's fleet of industrial mobile robots deployed across hundreds of customer sites. The company argues that this data-driven optimization approach produces more reliable and stable operations, with superior business logic and task execution algorithms compared to controllers that have not benefited from large-scale deployment experience.[1][2]
The VB1-I integrates what Lanxin calls a Multimodal AI Interaction Hub, a feature shared with the VB2. This system is designed to enable more natural communication between the robot and human operators or research subjects, supporting use cases in human-robot interaction research, natural language command interpretation, and multi-modal sensory integration studies.[12]
The following table compiles verified specifications from Lanxin's official product page and third-party robotics databases.[3][12]
| Category | Parameter | Value |
|---|---|---|
| Physical | Height | 1,650 mm (5 ft 5 in) |
| Physical | Dimensions (L x W x H) | 700 x 610 x 1,650 mm |
| Physical | Weight | 65 kg (143 lb) |
| Physical | Material | Aluminum alloy + composite |
| Physical | Ground clearance | 30 mm |
| Physical | Rotation diameter | 580 mm |
| Mobility | Driving mode | Dual-wheel differential motor |
| Mobility | Maximum speed | 0 to 1.5 m/s (5.4 km/h) |
| Mobility | Operating height range | 0.2 to 2.0 m |
| Manipulation | Arm joints | 6 per arm |
| Manipulation | Gripping load | 2 kg (per hand) |
| Manipulation | Fingers per hand | 5 |
| Manipulation | DOF per hand | 7 |
| Manipulation | Total degrees of freedom | 42 |
| Sensing | Navigation | 3D Laser SLAM |
| Sensing | Vision system | Panoramic RGB-D Depth Vision System |
| Actuators | Motor type | High-torque electric servo motors |
| Actuators | Gear technology | Harmonic reducers |
| Power | Battery life | 2 to 3 hours |
| Environment | Working temperature | 0 to 40 degrees Celsius |
| Environment | Working humidity | 10 to 90% RH (no condensation) |
| Computing | Operating system | Linux |
| Connectivity | Interfaces | Ethernet, WiFi |
| Safety | Safe with humans | Yes |
The VB1-I sits between the original VersaBot VB-1 and the VB2 in Lanxin's humanoid robot product line. Each model targets a different combination of use case, mobility, and capability level.
| Feature | VersaBot VB-1 | VB1-I | VB2 |
|---|---|---|---|
| Unveiled | August 2024 | 2024/2025 | 2025 |
| Mobility | Wheeled base | Dual-wheel differential | Four-wheel omnidirectional |
| Height | 1.6 m | 1.65 m | 1.7 m |
| Weight | Not disclosed | 65 kg | Not disclosed |
| Navigation | 3D Pure Vision (no LiDAR) | 3D Laser SLAM | 3D Laser SLAM |
| Arm joints | Not disclosed | 6 per arm | 7 per arm |
| Payload (per hand) | 2 kg | 2 kg | 5 kg |
| Fingers per hand | Gripper | 5 (dexterous) | 5 (dexterous) |
| Total DOF | Not disclosed | 42 | 47 |
| Battery life | Not disclosed | 2 to 3 hours | 4 to 6 hours |
| Maximum speed | Not disclosed | 1.5 m/s | 1.2 m/s |
| Operating height | 0.72 to 1.2 m | 0.2 to 2.0 m | 0.4 to 2.0 m |
| Price | Not disclosed | ~$120,000 | ~$150,000 |
| Primary focus | Industrial manufacturing | Research and embodied AI | Industrial and embodied AI |
| AI features | Core controller | Multimodal AI Interaction Hub | Multimodal AI Interaction Hub |
Several design trends are visible across the product line. Navigation shifted from pure vision (VB-1) to 3D Laser SLAM (VB1-I and VB2), suggesting that Lanxin found laser-based SLAM more effective for the precision requirements of humanoid robot applications. Hand design evolved from simple grippers (VB-1) to five-finger dexterous hands (VB1-I and VB2), reflecting the growing importance of manipulation research. Payload capacity and degrees of freedom increase from the VB1-I to the VB2, with the VB2 gaining an additional joint per arm (7 versus 6) and more than doubling the gripping load (5 kg versus 2 kg).[1][3][12][14]
The VB2's four-wheel omnidirectional drive provides greater maneuverability than the VB1-I's dual-wheel differential system, allowing holonomic motion (the ability to move in any direction without first rotating). The VB2 also offers significantly longer battery life at 4 to 6 hours versus 2 to 3 hours. However, the VB1-I is approximately $30,000 less expensive, making it a more accessible entry point for research teams with limited budgets.[12][14]
The VB1-I is designed primarily for research and development rather than continuous industrial deployment. Its target applications include:
The VB1-I provides a physical platform for researchers working on embodied artificial intelligence, the study of intelligent systems that learn through physical interaction with their environment. The robot's 42 degrees of freedom, five-finger dexterous hands, and multi-modal sensor suite allow researchers to experiment with reinforcement learning, imitation learning, and other AI training paradigms in a human-scale embodiment. The Linux-based operating system supports integration with popular AI frameworks and deep learning libraries.[3][12]
Research teams developing autonomous logistics solutions can use the VB1-I to prototype and test material handling workflows in laboratory settings before deploying them on production systems. The robot's 2 kg payload, variable operating height (0.2 to 2.0 m), and dual-wheel mobility allow it to simulate common factory tasks such as picking items from shelves, sorting components, and transporting lightweight parts between workstations.[3][12]
The VB1-I's human-scale form factor and Multimodal AI Interaction Hub make it suitable for studying how people respond to and collaborate with robotic systems. Researchers in cognitive science, ergonomics, and social robotics can use the platform to investigate questions about trust, collaboration patterns, and communication modalities in human-robot teams.[12]
The modular design of the VB1-I allows researchers to upgrade hardware components over time, making it suitable for long-term projects where the robot platform needs to evolve alongside the algorithms being developed on it. The standardized sensor suite and actuator system provide a consistent benchmarking environment for comparing different control strategies, path planning algorithms, and manipulation policies.[3]
In addition to its humanoid robot lineup, Lanxin Robotics has developed OmniHead, which the company describes as the world's first truly modular head for humanoid robots. Rather than requiring an entire system replacement when sensory or compute capabilities need upgrading, the OmniHead design allows operators to swap out individual components as needed. Lanxin has drawn an analogy to how smartphones became versatile through modular apps, arguing that humanoid robots will similarly become more adaptable through modular hardware.[15]
The OmniHead system is compatible with Lanxin's humanoid robot platforms and represents the company's broader strategy of building an upgradeable ecosystem rather than selling fixed-configuration robots.
The VB1-I entered the humanoid robotics market during a period of intense activity, particularly among Chinese companies. China's humanoid robot market is projected to reach approximately RMB 10.5 billion (US $1.4 billion) by 2026 and RMB 75 billion (US $10.3 billion) by 2029. Chinese firms dominated global humanoid robot shipments in 2025, with companies such as AgiBot (which shipped over 10,000 units by early 2026), Unitree Robotics, and UBTECH leading in volume.[16][17]
The research platform segment in which the VB1-I competes includes offerings from several international and Chinese manufacturers. Fourier Intelligence offers the GR-1 and GR-2 as research-oriented humanoids. EngineAI produces the SE01 and PM01 platforms targeted at developers and researchers. Internationally, Agility Robotics' Digit, 1X Technologies' NEO, and Unitree Robotics' G1 and H1 all serve overlapping research and early-industrial market segments.
Unitree's launch of its R1 humanoid in mid-2025 at a price of just $5,900 significantly disrupted pricing expectations across the industry, though the R1 targets a different capability tier than the VB1-I. The VB1-I's $120,000 price point positions it in the mid-range of the research humanoid market, below premium platforms from companies like Apptronik (Apollo) and Figure AI (Figure 02) but above the growing number of sub-$50,000 educational and entry-level platforms.[16][17]
Lanxin differentiates the VB1-I primarily through its heritage in industrial 3D vision technology. While most humanoid robot manufacturers began as robotics companies that later added vision systems, Lanxin started as a 3D visual perception company that expanded into robotics. This background gives Lanxin deep expertise in the sensor, algorithm, and software layers of the perception stack, which the company leverages across its entire product portfolio from standalone MRDVS sensors to integrated humanoid platforms.[4][6]
The VB1-I's use of a wheeled mobility platform places it in the "wheeled humanoid" category alongside robots such as the Enchanted Tools Mirokai, Pudu D9, and UniX AI Panther. This design approach trades the terrain versatility of bipedal locomotion for the stability, energy efficiency, and reliability that wheels provide in structured indoor settings. Wheeled humanoid robots have found growing adoption in retail assistance, customer service, indoor logistics, and research applications where navigating stairs or rough terrain is not required.[16][17]
The broader Chinese humanoid robot market includes both wheeled and bipedal platforms, with different companies choosing different approaches based on their target applications. Companies like AgiBot, Unitree, and UBTECH have focused primarily on bipedal designs, while Lanxin has consistently used wheeled bases across its VB-1, VB1-I, and VB2 platforms, reflecting its roots in industrial mobile robotics where wheels have proven more practical.[1][14]