Robbyant

• Robbyant R1

Supply chain relationships

• Hesai Technology

  Inbound - Perception - ROBBYANT R2 integrates Hesai JT-series lidar for indoor perception

  Source
• Orbbec

  Inbound - Perception - Robbyant (Ant Lingbo Technology) vision system uses an Orbbec camera

  Source

• Xing Zhu - CEO - Leads Robbyant's embodied-AI model and robot-platform strategy.
• Yujun Shen - Chief Scientist - Leads scientific work across Robbyant's embodied-AI model stack.
• Yongtao Huang - Chief Data Scientist - Leads data systems and training infrastructure for Robbyant's embodied-intelligence models.

• Robbyant open-sources LingBot-VLA model as a 'universal brain' for robots

  2026-01-28 - Research Publication - • Robbyant released LingBot-VLA, a vision-language-action model, as open-source software. • The model was pre-trained on over 20,000 hours of real-world interaction data from nine dual-arm robot configurations. • It was evaluated on the GM-100 benchmark, achieving higher task success rates than other models across three distinct robot platforms.

  Source
• Robbyant publishes LingBot-VLA pragmatic foundation model with open weights and benchmark data

  2026-01-26 - Research Publication - Robbyant published LingBot-VLA, a VLA foundation model trained on about 20,000 hours of real-world data from nine dual-arm robot configurations. The paper reports evaluations across three robotic platforms and releases code, base model weights, and GM-100 benchmark data for reproducible comparison. LingBot-VLA is a first-party Robbyant research release and one of the stronger open embodied-AI benchmark/data contributions in the humanoid-adjacent stack.

  Source

Events

• 2026-10-14 - Partnership - Robbyant and Dema Technology collaborate on flexible parcel sorting

  The logistics work combines model engineering, secondary development and validation of new sorting technologies. Robbyant’s CIIF demonstrations cover pharmacy picking, flexible logistics sorting and industrial loading/unloading. LingBot-VLA 2.0 ran on distinct Titanium Tiger and Astribot robot bodies during the exhibition.

  Source
• 2026-08-18 - Market Signal - Hesai discloses new humanoid lidar orders and actuation shipments

  New robotics-lidar orders and JT128 adoption include Unitree, Robbyant, Galbot, Galaxea and Dexmal. After Kosmo prototype deliveries in July 2026, orders were secured from humanoid-robot companies including Galbot, with initial revenue expected in Q3. The robotic-actuation production line is operational, dexterous-hand module shipments are underway and Sharpa is an actuation-module customer.

  Source
• 2026-07-24 - Capability - Hesai JT-series lidar equips ROBBYANT R2 indoor robot platform

  The ROBBYANT R2 platform from Ant Group’s Ant Lingbo embodied-intelligence company appeared at the 2026 World Artificial Intelligence Conference with Hesai JT-series lidar. The miniature hemispherical 3D lidar provides a 360° × 189° field of view and can be concealed within the robot while maintaining all-around detection, including close-range objects. The integration is positioned as a perception foundation for physical-AI robots, combining active probing, high-precision point-cloud modeling and interference resistance.

  Source
• 2026-07-17 - Demonstration - Orbbec and Robbyant demonstrate enhanced RGB-D data collection

  Orbbec’s EGO RGB-D line uses Gemini 330-series cameras for first-person physical-AI data capture. Robbyant adapted LingBot-Depth 2.0 to refine missing depth, object edges and spatial detail. The partners also demonstrated an enhanced-depth filter for commercial use with Gemini 330-series cameras.

  Source
• 2026-07-07 - Product Launch - Orbbec launched a robot-free data-collection hardware platform and announced model work with Robbyant

  Orbbec introduced dedicated hardware for robot-free data collection. The announcement also covered adaptation work with a Robbyant model. The adaptation was presented as collaboration rather than a completed standard deployment.

  Source
• 2026-05-26 - Partnership - GenRobot.AI partners with Ant Lingbo on human-behavior data

  The collaboration covers Ant Lingbo’s full embodied-model family. Joint development includes specialized capture devices and improvements to spatial accuracy, timing and action semantics. The programme targets real human interactions to improve cross-scenario model generalization.

  Source
• 2026-03-16 - Partnership - Robbyant and Leju Robot form strategic partnership to commercialize embodied AI systems

  Robbyant, Ant Group's embodied AI company, signed a strategic partnership with Leju Robot The two companies said they will combine Robbyant's embodied AI models with Leju's robotics technologies to build integrated solutions for industrial and commercial scenarios The partnership is focused on embodiment, data, and models, with the stated goal of accelerating embodied AI robots from specific task execution toward more general intelligence

  Source
• 2026-01-29 - Demonstration - Robbyant Demonstrates LingBot-VA's Video-Action Integration

  • Robbyant showcased LingBot-VA, an autoregressive diffusion framework that simultaneously learns frame prediction and policy execution. • The model features a shared latent space integrating vision and action tokens. • Evaluations demonstrated significant promise in long-horizon manipulation and strong generalizability to novel configurations.

  Source
• 2026-01-28 - Research Publication - Robbyant open-sources LingBot-VLA model as a 'universal brain' for robots

  • Robbyant released LingBot-VLA, a vision-language-action model, as open-source software. • The model was pre-trained on over 20,000 hours of real-world interaction data from nine dual-arm robot configurations. • It was evaluated on the GM-100 benchmark, achieving higher task success rates than other models across three distinct robot platforms.

  Source
• 2026-01-26 - Research Publication - Robbyant publishes LingBot-VLA pragmatic foundation model with open weights and benchmark data

  Robbyant published LingBot-VLA, a VLA foundation model trained on about 20,000 hours of real-world data from nine dual-arm robot configurations. The paper reports evaluations across three robotic platforms and releases code, base model weights, and GM-100 benchmark data for reproducible comparison. LingBot-VLA is a first-party Robbyant research release and one of the stronger open embodied-AI benchmark/data contributions in the humanoid-adjacent stack.

  Source

Relationships

• GenRobot.AI

  Human Data Collaboration - GenRobot.AI and Robbyant are jointly developing human-data production methods and coordinating data for the LingBot model family.

• Leju Robotics

  Partner - Robbyant and Leju partner to combine Robbyant AI with Leju humanoid robotics platforms.

• Dema Technology / 德马科技

  Technology Collaborator - Robbyant and Dema collaborate on high-volume flexible parcel sorting through engineering support, joint development and technology testing.