Company Overview
- Full Name: CloudMinds Robotics Co., Ltd. (达闼机器人股份有限公司)
- Founded: 2015
- Headquarters: Shanghai, China (global HQ), with R&D centers in Beijing, Shenzhen, and Chengdu
- Founder/CEO: Bill Huang (黄晓庆)
- Website: www.cloudminds.com
CloudMinds was founded in 2015 by Bill Huang, a legendary figure in China’s telecommunications industry. Huang previously served as Director of China Mobile Research Institute and CTO of UTStarcom, earning the nickname “Father of China’s Softswitch.” The company’s core philosophy is “cloud intelligence”—placing the robot’s “brain” in the cloud and connecting it to the robot body via high-speed networks. This architecture fundamentally differs from Tesla Optimus or Figure AI, which place all computing on the robot itself.
Core Products & Platforms
Cloud Robot Architecture
CloudMinds’ core technology is the “Cloud-Network-Edge” three-layer architecture:
- Cloud Brain: AI large models running in the cloud for perception, decision-making, and planning
- High-Speed Network: 5G/dedicated network for low-latency communication
- Robot Body: Terminal actuator receiving cloud instructions
HARIX OS
HARIX OS is CloudMinds’ proprietary cloud-native robot operating system and its core competitive moat:
- Manages thousands of robots simultaneously from the cloud brain
- Supports cloud-based robot skill deployment and real-time updates
- Provides a digital twin development environment
- Compatible with multi-brand, multi-form robot hardware
Robot Products
Ginger Series—Service Robots
- Ginger (Xiaojiang): CloudMinds’ flagship cloud service robot, ~160cm height, with facial recognition, voice interaction, and autonomous navigation. Designed for hotels, banks, shopping malls, and government halls. Priced ~$28,000-$42,000 depending on configuration
- Ginger Lite: Simplified version focused on reception and guided tours, ~$14,000-$21,000
XR+Robot Convergence
CloudMinds is pioneering XR (Extended Reality) integration with robots, enabling remote robot control through AR glasses/VR headsets:
- Remote inspection robots: Operators control robots in hazardous environments through XR devices
- Remote medical robots: Doctors perform medical procedures via XR-controlled robots
- Remote education robots: Experts guide robot manipulation remotely through XR
SCA (Smart Compliant Actuator)
CloudMinds’ proprietary flexible joint technology—a core motion control component integrating motor, reducer, encoder, and driver into a single unit with high torque density and force sensing capability.
API Platform & Developer Ecosystem
CloudMinds offers an open API platform and HARIX digital twin environment where developers can train robot skills virtually and deploy them to physical robots with one click.
Funding History
| Time | Round | Amount | Key Investors |
|---|---|---|---|
| 2016 | Series A | ~$30M | SBCVC, Gaorong Capital |
| 2017 | Series B | ~$100M | SoftBank Vision Fund |
| 2018 | Series C | ~$300M | SoftBank Vision Fund, Qiming Venture Partners |
| 2020 | Series D | ~$200M | China Internet Investment Fund (state-backed) |
| 2023 | New Round | ~$140M | Shanghai AI Industry Investment Fund |
CloudMinds originally planned a US IPO in 2021 but shelved it due to regulatory changes. It has since initiated STAR Market (Shanghai) IPO preparations. Total funding exceeds $1 billion, making it one of the most heavily funded robotics companies in China.
Competitive Positioning
Core Differentiation
- Architectural innovation: Centralizes computing in the cloud—reduces hardware cost and heat/power constraints on the robot body
- Remote brain: One cloud brain serves thousands of robots simultaneously; model updates happen in the cloud without hardware replacement
- HARIX OS ecosystem: Analogous to “Android for robots”; open platform attracts developers to build a robot skill app store
- SCA core technology: Among the few Chinese companies mastering proprietary robot flexible joints
Differentiation
- vs UBTECH/Unitree: Cloud architecture versus edge computing—fundamental technological divergence. Cloud offers greater flexibility but depends on network quality
- vs Tesla Optimus: Tesla puts all compute on the edge (end-to-end on-device AI); CloudMinds is purely cloud-based
- vs Beijing Humanoid Innovation Center: CloudMinds is a commercial company monetizing through HARIX OS; the Center is a public technology platform promoting open source
Industry Outlook
Prospects
Cloud architecture shows unique advantages in the large model era—models like GPT-5 and Pelican-VL are compute-hungry, and edge deployment faces cost and energy challenges. CloudMinds’ architecture naturally fits the “large models in the cloud” trend. The XR+Robot convergence direction offers commercial potential for remote operations, hazardous environment work, and industrial inspection.
Risks
- Network latency: Cloud robots require high-quality low-latency networks; performance suffers where 5G coverage is inadequate
- Commercialization pace: Service robot commercial deployment lags behind consumer products like robot vacuums
- Funding pressure: STAR Market IPO timeline is uncertain; ongoing capital needs are significant
- Competitive pressure: Edge computing is also advancing rapidly (Qualcomm, NVIDIA edge AI chips), narrowing the cloud-edge efficiency gap
Key Watchpoints
- Can HARIX OS attract enough third-party developers to form an ecosystem flywheel?
- Progress of the STAR Market IPO and post-IPO financial strength
- Will the cloud architecture be validated as the superior technical route in the embodied AI era?
- Can XR+Robot convergence find PMF in industrial inspection, remote healthcare, and other verticals?