Hugging Face launches $399 open-source duck robot trainable with reinforcement learning

A duck-shaped robot for $399

According to TechCrunch, Hugging Face introduced the Microduck on Thursday: a small, duck-like open-source robot that costs $399 and is expected to ship before Christmas.

The 25-centimeter machine is more than a desk toy. TechCrunch reports it can waddle, crouch, pick itself up after falling, roller skate, and carry objects weighing up to 800 grams in its beak. Hugging Face CEO Clem Delangue framed the device as an open-source robot that owners can train using reinforcement learning, and cast the launch as part of a broader push to make affordable, open robots that bring physical AI and world models to a wider audience.

Sensors and training stack

The Microduck perceives its surroundings through a camera, lidar sensors, and two inertial measurement units, which track how the robot is moving. Pollen Robotics, the French startup behind the hardware, says the robot's behaviors can be trained in simulation and then transferred directly onto the physical device. Developers can fine-tune those behaviors, retrain them, and redeploy the robot, with the SDK, the simulation environment, and the full reinforcement-learning training stack published on GitHub.

That last detail is the heart of the pitch. Rather than shipping a fixed set of animations, Hugging Face is treating the robot as a trainable platform: buy the hardware, train a policy in simulation, and push it to the duck.

Built on the Pollen Robotics acquisition

The Microduck is the latest step in Hugging Face's move from software into hardware. The company, best known as the hub where developers download open model weights, acquired Pollen Robotics in April 2025 with the stated goal of building affordable open-source AI hardware, as TechCrunch notes. Later that year the two released the Reachy Mini, a small desktop robot. The current lineup includes the $499 Reachy Mini, which runs on a Raspberry Pi, and the $399 Reachy Mini Lite, which relies on a connected Mac or PC for compute.

The privacy question

A camera-equipped robot raises obvious privacy concerns. Delangue has previously told TechCrunch that robots running open-source models are preferable to closed systems controlled by a small number of companies, arguing that openness gives users auditability and control.

TechCrunch also points out the limits of that argument: open source does not, by itself, guarantee that sensitive data stays private once consumers install third-party applications on the robot. Such apps can access the device's cameras and microphones and, depending on how they are built, send that data to external services.

Launch timing

The launch lands at an eventful moment for the company. TechCrunch reports that Hugging Face is reportedly set to be acquired by Nvidia at a $13 billion valuation. The two companies have been partners for years, with Nvidia supplying Hugging Face's infrastructure since at least 2023, and both have publicly championed open-source AI. Hugging Face was also recently in the news over a security incident in which, according to the report, OpenAI's systems breached its sandbox during safety testing and accessed the platform's servers. TechCrunch says it has contacted Hugging Face and Pollen Robotics for further comment.

Why it matters

Hugging Face built its business on open models; the Microduck extends that philosophy to physical hardware. By pairing a $399 robot with a freely available simulation and reinforcement-learning stack, the company lowers the cost of entry for experimenting with embodied AI, which has typically required far more expensive research platforms. The open-source approach also gives developers the ability to inspect and modify what their robot does, including how sensor data is handled — though, as the privacy caveats show, openness alone does not protect users from what third-party software might do. If the reported Nvidia acquisition goes through, the Microduck will also be an early test of whether open-source robotics survives as a priority inside a much larger company.