What Figure AI does
Figure AI is building general-purpose humanoid robots intended to operate in human environments, starting with “workforce” deployments and expanding toward home use. Its core approach couples embodied hardware (Figure 01/02/03) with an AI “vision-language-action” stack called Helix, designed to control whole-body locomotion and manipulation in cluttered, changing spaces. Figure is also pursuing scale through vertical integration of manufacturing, including its high-volume facility BotQ, and through large-scale data/compute programs that it says address key bottlenecks for training generalist physical intelligence. The company’s published strategy emphasizes shipping production-ready humanoids, improving reliability and manufacturing yield, and growing fleet-generated data so Helix can learn and generalize across many environments and tasks.
Figure’s most visible products/platforms are: (1) its humanoid robot generations (Figure 01, Figure 02, Figure 03); (2) Helix, which it positions as a generalist humanoid “Vision-Language-Action” model powering autonomous control; (3) BotQ, its in-house manufacturing facility meant for high-volume output; and (4) Index, described as a pipeline for collecting and processing large-scale physical training data. In parallel with product development, Figure has announced collaborations/partnerships aimed at expanding compute capacity (Nscale on NVIDIA Vera Rubin) and scaling real-world data collection and training environments (Brookfield partnership).