Microduck

A tiny open-source biped you can train yourself

RobotsArtificial IntelligenceOpen Source

▲ 224 votes6 commentsLaunched Aug 28, 2026

Visit Website

Daily #1Weekly #20

Microduck is a 25cm, $399 open-source bipedal robot built by Hugging Face and Pollen Robotics. It’s designed for sim-to-real reinforcement learning. It ships with 7 pre-trained behaviors and an Apache 2.0 software stack you can clone, modify, and retrain.

AI Analysis

📝 Summary

Microduck is a 25cm tall, $399 open-source bipedal robot developed by Hugging Face and Pollen Robotics for sim-to-real reinforcement learning. It features 7 pre-trained behaviors and a fully modifiable Apache 2.0 software stack, allowing users to clone, tweak, and retrain AI models. Unique selling points include its affordability, open hardware/software, and focus on accessible embodied AI experimentation. It solves key pain points like the high cost and closed nature of traditional robotics platforms, enabling researchers, educators, and hobbyists to innovate without prohibitive barriers. The value proposition is democratizing advanced robotics and RL research through an easy-to-use, trainable platform.

📈 Market Timing

In 2025-2026, market timing is favorable due to booming embodied AI, maturing sim-to-real RL tech, and rising demand for affordable open-source robotics amid AI democratization trends. Economic support for tech innovation and open-source movements further align. Excellent Timing as it capitalizes on the shift toward practical, trainable robotics beyond expensive industrial solutions.

✅ Feasibility

High. Technical difficulty is mitigated by pre-trained models and open-source design; development costs are controlled at $399 price point with established partners. Supply chain for compact robots is feasible but scaling manufacturing carries moderate risks. Strong scalability in research/education sectors with good team backing from Hugging Face and Pollen Robotics.

🎯 Target Market

Primary segments: AI/robotics researchers and students (academia/industry), STEM educators, open-source hardware hobbyists. Demographics: tech-savvy individuals aged 20-45, mainly in North America, Europe, East Asia. TAM for AI robotics education/research ~$2B, SAM for small bipeds ~$150M, SOM ~$10M. Core pain points: lack of cheap, modifiable platforms for RL experiments. High willingness to pay among research budgets; $399 appeals broadly.

⚔️ Competition

Medium. Direct competitors: 1. NAO (Softbank Robotics, https://www.softbankrobotics.com/), 2. Poppy Humanoid (Poppy Project, https://www.poppy-project.org/), 3. Unitree H1 (Unitree Robotics, https://www.unitree.com/), 4. iCub (Italian Institute of Technology, https://icub.iit.it/). Advantages: far lower price, fully open-source with pre-trained RL behaviors and Hugging Face integration. Disadvantages: smaller scale may limit some real-world applications vs larger, more robust competitors; less brand recognition in education.

Upgrade Pro to unlock full AI analysis