Figure AI releases a new generation of humanoid robots: smoother movements and more dexterous hands
B.news
Figure AI, an American humanoid robotics startup, officially launched its third-generation humanoid robot, Figure 03, on its official website on Thursday, October 9th. In a 6-minute-and-19-second demonstration video, Figure 03 seamlessly completed a series of daily tasks, including serving tea and water, interacting with pets, tidying the room, washing dishes, loading laundry, and other household tasks.
It also demonstrated its ability to perform tasks in a variety of scenarios, including hotel reception, express delivery, package sorting, and even banquet service. Compared to the previous generation, Figure 03's movements are more natural and fluid, significantly improving overall performance.
The robot's exterior is covered in fabric, replacing traditional metal structures. According to the company, this soft covering is easily removable, replaceable, and washable, and can be matched with a variety of clothing, enhancing its adaptability. Furthermore, Figure 03's hands are equipped with wide-angle, low-latency cameras, providing close-range visual feedback even when the main camera is obscured, such as when reaching into a closet or a narrow space, enabling continuous, real-time adaptive control. The company also revealed that the robot's feet are equipped with built-in induction coils, allowing it to be charged at 2kW via a wireless charging dock. At home, Figure 03 can automatically return to its charging station, recharge itself, and simultaneously upload data.
Figure AI emphasized that Figure 03 is the first humanoid robot designed from the ground up for mass production, with nearly all components redesigned to simplify assembly, reduce the number of parts, and lower costs.
Figure 03's environmental adaptability marks a fundamental upgrade in its AI core. It no longer simply executes preset, isolated commands, but instead autonomously plans and makes decisions in a dynamic, unstructured, real-world environment. Auxiliary cameras built into its palms and a fully upgraded vision system form a comprehensive perception network, enabling it to maintain awareness even when its field of vision is partially obstructed. This capability is the prerequisite for the robot's transformation from a laboratory showcase to a truly universal workforce capable of independent operation in complex environments such as homes, warehouses, and hotels.
Figure AI's announcement that Figure 03 is the "first robot designed for high-volume manufacturing" has perhaps a more profound significance than any technical specification. It signals a shift in the industry's competitive focus from "can you build one powerful robot" to "can you build thousands of affordable and reliable robots." When companies begin to streamline assembly, reduce parts, and lower costs, it signals the dawn of technology commercialization, and the threshold for large-scale application is about to be crossed. Throughout technological development, artificial intelligence once resembled a torrent surging through the digital world.
The emergence of Figure 03 is like finding a channel for this torrent to flow into the physical world. It demonstrates that the next wave of AI will not only be about better algorithms and larger models, but also about embedding this intelligence into a body capable of action, perception, and interaction. This giant ship, known as "embodied intelligence," has sounded its whistle and is slowly leaving the harbor of concepts. Its destination will be the real world that connects us all.
