Specifications
Manufacturer: Figure AI
Height: 172 cm (5'8")
Weight: 61 kg (134 lb)
Degrees of Freedom: 35, including human-like wrists, hands, and fingers
Payload Capacity: 20 kg (44 lb)
Battery: Approximately 5 hours, with 2 kW fast charging
Sensors: Embedded palm cameras, custom tactile sensors (3-gram force resolution), wide-FOV camera system with 2x frame rate vs. Figure 02
Hands: Dexterous multi-finger with tactile sensing at each fingertip
Key Improvement Over Figure 02: 9% less mass, significantly reduced volume, 60% wider camera field of view
Status: Pre-production, commercial deployment planned
Current Status
Figure 03 represents a significant refinement from Figure 02 rather than a ground-up redesign. The 9% weight reduction and volume decrease come from engineering optimization: tighter packaging of actuators, more efficient battery chemistry, and improved structural design.
The sensor upgrades are substantial. The camera architecture now delivers twice the frame rate and one-quarter the latency of Figure 02, with a 60% wider field of view per camera. The embedded palm cameras give Figure 03 close-up visual intelligence that most humanoids lack entirely, allowing it to inspect what it holds and manipulate small objects with visual guidance.
The custom tactile sensors can detect forces as small as 3 grams, which is roughly the weight of a single sheet of paper. This level of sensitivity opens applications that require delicate handling: assembling electronics, folding laundry, or handling fragile items in retail environments.
Clothing Considerations
Figure 03 is the benchmark for robot garment design. Figure AI's decision to make soft goods a core system means the platform was engineered with garment interfaces from the start, not retrofitted.
The knitted covering. Figure 03's standard soft goods use a knitted textile that stretches across joints, conforms to the body during movement, and recovers its shape at rest. The material choice (knit over woven) was deliberate: knit fabrics have inherent multi-directional stretch that accommodates the robot's kinematics without restricting motion or developing stress tears.
Tool-free removal. The covering system is designed for removal without tools. This is significant for maintenance (the robot body needs periodic inspection and cleaning) and for garment replacement. A fleet operator can swap coverings in minutes rather than hours.
Machine washable. The soft goods survive standard machine washing. This sounds simple but required careful material selection: the fabric, closures, and attachment hardware all had to withstand repeated wash cycles without degrading fit or function.
Low-friction liner. Between the knitted covering and the robot's body surface sits a low-friction liner layer. This solves the problem that plagues garments on robots with silicone or rubber skin: high-friction contact that makes clothing nearly impossible to slide on and off. The liner lets the covering glide into place and remove cleanly.
Fashion Potential
Figure 03 is the platform most likely to attract fashion designers rather than just garment engineers. The clothing interface is clean, well-documented, and designed for third-party use. A designer working with Figure 03 spends their time on aesthetics and function, not fighting the platform.
The 35-DOF body creates a natural, fluid movement profile that makes garments look good in motion, not just static display. This matters for any application where the robot is customer-facing: hospitality, retail, healthcare, reception.
Third-party garment programs for Figure 03 will likely be the first to reach commercial scale. The combination of a garment-friendly platform, clear attachment standards, and growing commercial deployment creates the conditions for an aftermarket garment ecosystem to develop.
For a deep dive into Figure 03's textile engineering, see our article on Figure 03's Washable Textiles. For a broader view of materials that work on robots, see our materials guide.