Figure AI Inc.
Company Overview
Figure AI Inc. is an American robotics and artificial intelligence company headquartered in Sunnyvale, California. Founded in 2022 by serial entrepreneur Brett Adcock (formerly of Archer Aviation and Vettery), the company is dedicated to developing autonomous general-purpose humanoid robots. Unlike traditional single-purpose industrial robots, Figure's machines are designed with a human form factor to operate seamlessly in environments built for humans, such as factories, warehouses, and eventually homes.
The company has rapidly achieved "unicorn" status, backed by high-profile investors including Jeff Bezos, Nvidia, Microsoft, OpenAI Startup Fund, and Intel Capital. Figure AI distinguishes itself through its "Master Plan," which outlines a roadmap from developing a commercially viable humanoid to achieving mass-market household adoption.
Continue…Core Mission
Figure's primary mission is to deploy autonomous humanoid robots to address global labor shortages and reduce the need for humans to perform unsafe, undesirable, or repetitive jobs. The company operates on the philosophy that because the physical world is designed for humans (door handles, stairs, tools, aisles), a humanoid form factor is the most efficient design for a general-purpose robot.
Products
Figure 02 (Current Flagship)
Released in August 2024, Figure 02 is the company's second-generation humanoid and its first commercially viable product designed for mass production and real-world industrial deployment. It represents a significant hardware and software overhaul from the previous prototype.
- Form Factor: 5'6" (1.68m) height, approximately 70kg weight.
- Payload: 20kg (44 lbs) capacity.
- Runtime: Approximately 5 hours per charge using a custom 2.25 kWh battery pack integrated into the torso.
- Locomotion: Bipedal walking speed of 1.2 m/s with human-like gait and balance.
- Design: Unlike its predecessor, Figure 02 features a matte black "exoskeleton" finish with integrated cabling (no exposed wires), making it safe for tight workspaces.
Key Hardware Features:
* Dexterous Manipulation: The robot features fourth-generation hands with 16 degrees of freedom (DoF) and five fingers, equivalent to human hand size and strength. This allows it to grasp complex objects, use tools, and perform fine motor tasks.
* Vision System: Equipped with six onboard RGB cameras driven by an AI-powered vision system, allowing the robot to build a 3D understanding of its environment.
* Compute: Onboard computing provides 3x the inference power of the previous generation, enabling fully autonomous operation without reliance on remote tethers.
Figure 01 (Prototype)
Figure 01 was the company's proof-of-concept robot, unveiled in 2023. It served as the testbed for the company's end-to-end AI models.
* Achievements: demonstrated the ability to walk, make coffee autonomously (using a Keurig machine), and perform warehouse bin-moving tasks.
* Legacy: While successful, it had exposed wiring and a bulkier aluminum chassis, which were refined for the production-ready Figure 02.
Technology & Artificial Intelligence
Figure AI utilizes an "End-to-End" Neural Network approach. Rather than relying on hard-coded programming for every movement, the robots learn tasks through observation and simulation.
OpenAI Partnership & Speech-to-Speech
A defining feature of Figure's technology is its collaboration with OpenAI. This partnership integrates high-level AI models directly into the robot, providing:
* Speech-to-Speech Reasoning: Operators can speak to the robot in natural language. The robot processes the audio, "thinks" (plans a response/action), and speaks back while simultaneously executing physical tasks.
* Semantic Understanding: The robot can interpret vague commands (e.g., "I'm hungry") and translate them into logical physical actions (e.g., handing the human an apple from a counter) without explicit step-by-step coding.
Vision Language Models (VLM)
The robots utilize onboard Vision Language Models to map visual inputs to actions. This allows the robot to identify objects, assess whether they are trash or valuable items, and determine how to manipulate them based on context.
Strategic Partnerships and Commercialization
BMW Manufacturing
In January 2024, Figure AI signed a major commercial agreement with BMW Manufacturing.
* Deployment: Figure robots are being deployed at BMW's Spartanburg, South Carolina plant?the largest automotive exporter in the U.S.
* Application: The robots are tasked with identifying and placing sheet metal parts and performing chassis assembly tasks. This partnership serves as the primary validation ground for Figure's hardware in a high-stakes industrial environment.
Microsoft Azure
Figure utilizes Microsoft Azure for AI infrastructure, training, and storage. This partnership allows Figure to access massive compute resources needed to train the large-scale AI models that control the robots' perception and action loops.
Nvidia
Figure utilizes Nvidia's GPU technology for both onboard inference (processing data on the robot) and simulation training (learning in a virtual environment before attempting tasks in the real world).
