Story
August 29, 2026

Hugging Face Thinks a Cute $399 Duck Can Crack Consumer Robotics

Hugging Face’s Microduck pairs open-source AI training with a playful design, betting that consumers will embrace a robot built to wobble, fall and learn. Early sales excitement is strong, but its usefulness remains deliberately modest.

Hugging Face is betting that robotics becomes less intimidating when it arrives as a tiny, roller-skating duck. The company sees Microduck as an affordable gateway to hands-on AI learning; skeptics need not look far to see a cute novelty with limited household utility.

On Thursday, Hugging Face unveiled the $399 Microduck, a one-eyed, roughly 10-inch robot developed through its Pollen Robotics unit. It can follow a laser pointer, pick up small items with its beak-like gripper, skate, and recover after a tumble; it also carries a camera, LiDAR and motion sensors. Preorders are open in four colors, with first deliveries planned before Christmas 2026.

The pitch is not that the duck will take over the chores. Axios described it as “cute-but-limited,” noting its sub-two-pound frame and narrow practical ambitions. Instead, its open-source software is meant to let developers train new behaviours through reinforcement learning — trial and error in miniature.

That embrace of imperfection is central to Hugging Face CEO Clem Delangue’s argument. Microduck was designed to be “made to move, ready to fall,” he said, adding that people should understand “that robots can fail and make mistakes.” In his launch post, Delangue framed the machine as a teachable open-source robot that can walk, pick things up, regain its footing and roller-skate.

The response suggests the toy-like design may be the feature, not the distraction. One widely shared post claimed that “Thousands of microducks” sold on the first day, concluding: “The people want a cute robot.” Another reported more than $1 million in sales.

Delangue has set an ambitious benchmark: 50,000 units would make Microduck a major business, or about $20 million in revenue. The harder test comes after the first rush of quacks — whether buyers keep training the duck once its charm has worn off.