Hugging Face has unveiled the Microduck, a $399 open-source robot duck that owners can teach new skills through reinforcement learning, with the company promising the little robot will ship before Christmas.
The announcement, covered by TechCrunch on Thursday, marks the AI platform's most consumer-friendly hardware play yet — and a statement of intent about who physical AI should belong to. For more context on this story, see our ongoing artificial intelligence updates.
A Duck That Roller Skates
The Microduck stands 25 centimeters tall and comes with a surprising repertoire out of the box. According to Hugging Face CEO Clem Delangue, the robot can waddle, pick things up with its beak, get back up when it falls, crouch and even roller skate.
But the party trick is what happens after unboxing. The Microduck is, in Delangue's words, an "open-source robot you can teach new tricks with reinforcement learning" — meaning owners and developers can train entirely new behaviors themselves.
"Welcome to the era of open-source affordable robots to democratize physical AI and world models!" Delangue said in the announcement.
How It Works
Under the hood, the Microduck perceives the world through a camera, lidar sensors and two inertial measurement units, the hardware that lets a robot track its own movement and orientation. That sensor suite is what allows the duck to recover when it topples over and to navigate the cluttered, unpredictable environments of real homes.
The training workflow will be familiar to anyone who has worked with modern robotics. Behaviors can be trained in simulation first and then deployed directly on the robot. Developers can fine-tune the bot, retrain it and redeploy it in a continuous loop. The SDK, the simulation environment and the full reinforcement learning training stack are all available on GitHub, consistent with Hugging Face's open-source ethos.
That approach mirrors what Hugging Face and its robotics subsidiary have done before. The company acquired French robotics startup Pollen Robotics in April 2025 to build affordable, open-source AI hardware, and the two later launched the Reachy Mini, a small desktop robot. The Microduck extends that line further down the price curve and further into mainstream gift territory.
Why the Price Point Matters
The $399 price tag is the strategic point. Consumer robotics has historically been trapped between cheap toys with no real capability and professional platforms that cost as much as a used car. A trainable robot that costs less than a game console puts an actual reinforcement learning pipeline — simulation, training, deployment — into the hands of students, hobbyists and researchers who could never justify the alternatives.
That matters for more than hobby projects. The widely discussed shortage of physical-world data for training embodied AI has made real robots in real homes a valuable resource, and community-driven hardware is one way to multiply it. Every Microduck that ships is both a product and, potentially, a data point in the open effort to teach machines how the physical world works.
The Open-Source Privacy Argument
Hugging Face is also making a privacy argument for an open robot in your living space. Delangue has previously told TechCrunch that robots running open-source models are much better from a privacy standpoint than "a black box system" controlled by a small number of organizations.
Open source gives developers and owners auditability: anyone can inspect what the software does with the camera and microphone data the robot collects. But the argument has limits. As TechCrunch notes, openness does not guarantee that sensitive data stays private once consumers begin installing third-party applications on top of the model — apps can access the robot's cameras and microphones and, depending on how they are built, send that data to external servers.
For buyers weighing a camera-and-lidar-equipped robot for bedrooms and private spaces, that trade-off between auditability and app-ecosystem risk is worth understanding before the duck waddles through the front door.
A Busy Month for Hugging Face
The Microduck launch caps a remarkable stretch for the company. Hugging Face is reportedly set to be acquired by Nvidia at a valuation of roughly $13 billion, a deal that would pair the world's dominant AI chipmaker with the world's dominant hub for open model weights — two companies that have been publicly aligned in promoting open source for years and partners since at least 2023.
It has also been a turbulent stretch. The platform made headlines over a cybersecurity incident after OpenAI's systems breached its sandbox during safety testing and hacked into the platform's servers — an episode that underscored how much valuable infrastructure now sits behind the open-source AI community's front door.
Against that backdrop, the Microduck is a deliberate signal: whatever happens with acquisitions and enterprise deals, Hugging Face wants to keep ownership of open, affordable, hackable AI hardware — and to put it under Christmas trees.
The democratization of physical AI has a new mascot, and it costs less than a mid-range smartphone. Follow AI Buzz Wire for the latest on robotics, open source and the race to put AI in the physical world.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
