Agile ONE is the first humanoid robot by Agile Robots SE (Munich, Germany), launched in November 2025. Standing 174 cm and weighing 69 kg, it features 21 DOF per hand with fingertip and force-torque sensors, AI-layered architecture with Robotic Foundation Models, and targets industrial material handling, machine tending, and fine manipulation at up to 2 m/s walking speed.
Primary use case: Industrial manufacturing, material handling, assembly, machine tending
Height
174 cm
Weight
69 kg
DOF
42
Payload
20 kg
Battery
8 hrs
Height
174 cm
Weight
69 kg
Degrees of Freedom
42
Locomotion
bipedal
Hand Type
5-Finger Dexterous
Payload
20 kg
Max Lift/Arm
10 kg
Battery Life
8 hrs
Autonomy Level
Semi-Autonomous
Availability
In Development
Primary Use Case
Industrial manufacturing, material handling, assembly, machine tending
Audi is testing Agile Robots’ mobile humanoids at its Ingolstadt Production Lab to retrieve varied, flexible and unsorted automotive components from containers and sequence them for assembly. Announced on October 21 at NVIDIA GTC Berlin, the proof of concept involves Dassault Systèmes and NVIDIA virtual testing but has yet to disclose quantified performance results or a factory-wide rollout schedule.
Agile Robots launched the Agile ONE SN at NVIDIA GTC Berlin, introducing a 1.4-metre, 43-kilogram research humanoid with 71 degrees of freedom, tactile hands, Jetson Thor hardware and the GR00T 1.7 VLA model. The company also introduced the Diana 7 and Diana 3 arms and joined AUDI AG, Dassault Systèmes and NVIDIA on a humanoid logistics pilot at Audi's Ingolstadt plant. NVIDIA separately announced plans to invest in Agile Robots, although the amount was not disclosed.
VicOne reports that AI-assisted research uncovered 28 initially validated zero-day vulnerabilities in less than two days across software and tools supporting robotics development, training, simulation, deployment, and runtime, including projects maintained by Google, NVIDIA, and Meta. The article warns that compromised components could affect models, test results, operator visibility, remote-stop functionality, and robot behavior, and recommends identifying affected systems, assessing potential consequences, and validating patches and temporary safeguards.