Alibaba Group
About
About Alibaba Group
Qwen’s robotics work separates perception and language reasoning from lower-level action generation. Qwen-VLA combines a Qwen language-and-vision model with a diffusion-based action component, allowing one architecture to interpret instructions and produce manipulation trajectories while retaining information about the target robot embodiment.
The wider Qwen-Robot work includes models and tools for manipulation, navigation, world modelling and simulation transfer. Alibaba releases model weights, code and developer documentation rather than manufacturing a standard robot body, allowing laboratories and platform builders to adapt the software to different arms, mobile bases and simulated environments.
Supply
Supply chain relationships
- SLAMTEC
Inbound - Navigation - Alibaba Diana inventory robot used SLAMTEC SLAMWARE and a 40 metre dTOF lidar
Source - Covestro
Inbound - materials and robotics supply chain - Covestro’s mass-balance-attributed Makrolon RE was selected for Alibaba DAMO’s Xiaomanlv power-grid inspection robot
Source - RoboSense
Inbound - Perception - Alibaba Xiaomanlv logistics robot uses RoboSense LiDAR
Source
Research
- Amap releases open ABot-Recon streaming 3D model
2026-08-28 - Research Publication - ABot-Recon accepts monocular RGB video without requiring depth sensors or pre-calibrated camera parameters. On KITTI-02 at 504 by 280 resolution, ABot-Recon reaches 24.45 frames per second with about 6.71 GB peak memory; the stated minimum GPU is a GTX 1080 Ti. Inference code, evaluation scripts and pretrained weights were public at announcement; no future training-code release is treated as completed.
Source - Alibaba AMAP releases ABot-World-0 interactive world model
2026-07-21 - Research Publication - The model produces 720p interactive video at up to 16 FPS with about 1.2 seconds of latency and roughly 19 GiB peak VRAM on an RTX 5090. ABot-World-0 accepts raw keyboard actions for roaming and character interaction rather than relying on a fixed scripted trajectory. The release includes a technical report, project page, public code and downloadable model weights.
Source - Alibaba releases Qwen-Robot embodied-AI model suite
2026-06-16 - Research Publication - The suite includes Qwen-RobotNav, Qwen-RobotManip and Qwen-RobotWorld components. The release targets robot navigation, manipulation policy development and simulation workflows.
Source
Timeline
Events
- 2026-08-28 - Research Publication - Amap releases open ABot-Recon streaming 3D model
ABot-Recon accepts monocular RGB video without requiring depth sensors or pre-calibrated camera parameters. On KITTI-02 at 504 by 280 resolution, ABot-Recon reaches 24.45 frames per second with about 6.71 GB peak memory; the stated minimum GPU is a GTX 1080 Ti. Inference code, evaluation scripts and pretrained weights were public at announcement; no future training-code release is treated as completed.
Source - 2026-08-24 - Funding - XPENG Robotics signs share-purchase agreements for more than $900M
IDG Capital initiated the round, with Gaorong Ventures participating and Tencent and Alibaba supporting as strategic investors. The capital is allocated to robotics hardware and software R&D, Physical AI model training, data generation, production facilities and international expansion. The announcement concerns signed financing agreements; it does not establish that every subscription had closed on August 24, 2026.
Source - 2026-07-21 - Research Publication - Alibaba AMAP releases ABot-World-0 interactive world model
The model produces 720p interactive video at up to 16 FPS with about 1.2 seconds of latency and roughly 19 GiB peak VRAM on an RTX 5090. ABot-World-0 accepts raw keyboard actions for roaming and character interaction rather than relying on a fixed scripted trajectory. The release includes a technical report, project page, public code and downloadable model weights.
Source - 2026-06-16 - Research Publication - Alibaba releases Qwen-Robot embodied-AI model suite
The suite includes Qwen-RobotNav, Qwen-RobotManip and Qwen-RobotWorld components. The release targets robot navigation, manipulation policy development and simulation workflows.
Source - 2017-07-13 - Deployment - Geek+ robots operate at Alibaba’s Tmall warehouse
Alibaba’s Tmall warehouse was the named operating site. Geek+ said dozens of robots were working in the warehouse. Geek+ reported that the warehouse use had begun in 2015.
Source
Context
Comparable profiles
- ACE Robotics (大晓机器人)
comparable profile - AI Robot Model Framework
- RoboScience / 机器科学
stack adjacency - Robotics Foundation Model Labs
- XYZ Embodied AI
stack adjacency - Embodied Foundation Model Developer
- Lupa Robotics
stack adjacency - Embodied Foundation Model Developer · Robot Foundation Model Platform
- ForgeAI
stack adjacency - Embodied Foundation Model Developer · Robot Foundation Model Platform
- Sereact
stack adjacency - Robot Foundation Model Platform
- Physical Intelligence
stack adjacency - Robot Foundation Model Platform
- Aether AI
stack adjacency - Foundation Model
- One Robot
stack adjacency - Robotics Foundation Model Labs
- General Trajectory
stack adjacency - Robotics Foundation Model Labs
- Tommoro Robotics
stack adjacency - Robot Foundation Model Platform
- AstraQua
stack adjacency - Foundation Model
Context tags
- Robotics foundation-model ecosystem - Stack Layer - https://autonews.gasgoo.com/articles/news/alibaba-launches-qwen-robot-series-embodied-ai-large-models-2066890758655971328
- Foundation model - Stack Layer - https://qwen.ai/blog?id=qwen-robotsuite
- Embodied AI - Capability
- Robot foundation model platform - Stack Layer - https://qwen.ai/blog?id=qwen-robotsuite
- Embodied foundation model developer - Stack Layer - https://qwen.ai/blog?id=qwen-robotsuite
- AI robot model framework - Peer Group - https://autonews.gasgoo.com/articles/news/alibaba-launches-qwen-robot-series-embodied-ai-large-models-2066890758655971328