Humanoid Database
Disclaimer: All content presented on this site, including robot profiles, images, and descriptions, is provided for editorial, educational and informational purposes only. Humanoid Press does not claim ownership of any third-party trademarks, designs, or intellectual property featured herein. All rights belong to their respective owners.
GR-2: Fourier Intelligence
|
type: |
type 1 |
|
{{variant.name}}:
|
{{opt.name}}
{{opt.name}}
|
Summary:
GR-2 is Fourier Intelligence’s second-generation humanoid platform: a 1.75-meter, ≈63-kg robot designed for real-world tasks, with 53 degrees of freedom, high-torque actuators, and 12-DoF dexterous hands equipped with tactile sensing. Launched in 2024, GR-2 focuses on robustness, modularity, and developer tooling for industrial and research applications.
Editorial:
When Fourier Intelligence announced GR-2, the company’s message was clear: this is a humanoid built for work, not spectacle. Rather than chase purely aesthetic milestones, GR-2 doubles down on practical upgrades — stronger actuators, more refined hands, and software that helps move research from simulation to the factory floor.
Fourier lists GR-2 at roughly 1.75 m tall and 63 kg, with a 53-degree-of-freedom body and 12-DoF dexterous hands outfitted with tactile sensor arrays. Those hands, together with the robot’s upgraded actuators, are intended to give GR-2 far greater manipulation fidelity than many earlier humanoids — useful when a task requires delicate gripping as well as brute force.
Hardware matters here. Fourier’s documentation highlights new FSA-series actuators and a higher torque envelope, allowing the platform to handle heavier tasks while keeping response times tight. The GR-2 also uses integrated cabling and a modular layout to shorten maintenance cycles — a design choice that speaks directly to customers in manufacturing and field service who need uptime and repairability.
Beyond raw mechanics, Fourier positions GR-2 as an ecosystem: an SDK and support for common robotics frameworks (ROS, NVIDIA Isaac, Mujoco) that let developers translate AI models into physical behavior more quickly. That software layer is as important as the metal and motors — it determines whether a lab prototype can become something useful on a real shop floor.
Early demos and coverage show GR-2 walking, manipulating objects, and operating in human environments — a necessary but still partial demonstration. Real-world deployment will hinge on long-term reliability, battery life in continuous operation, and safety certification in industrial settings. Fourier reports a swappable battery design that extends runtime, which helps but does not eliminate those operational questions.
Strategically, GR-2 is a statement: companies in China are not only building humanoids to showcase onstage, they are solving the messy engineering problems that follow — maintenance, integration, power systems, and developer tooling. For buyers and researchers, that shift makes GR-2 interesting because it’s designed to be used, not just admired.
Image: humanoidrobotsguidetothegalaxy