Agile Robots to Build 2,000 Humanoids With NVIDIA Tech
Munich-based Agile Robots plans to produce up to 2,000 humanoid robots in 2027, all powered by NVIDIA technology. CEO Zhaopeng Chen shared with the media that the company will begin building the robots by the end of 2026.
The company unveiled its new Agile ONE SN robot on October 21 at NVIDIA’s GTC conference in Berlin. The “SN” stands for the NVIDIA-specific version of the robot. It runs on NVIDIA’s Jetson Thor computing platform and uses the Isaac GR00T AI model for learning and movement.
The Agile ONE SN stands 1.4 meters tall and weighs about 40 kilograms. It has 31 degrees of freedom, five-fingered hands, and built-in cameras. As a result, the robot can handle precise tasks like grabbing objects and interacting with its surroundings.
AGILE ROBOTS UNVEILS NVIDIA-POWERED HUMANOID, TARGETS UP TO 2,000 ROBOTS IN 2027
Germany’s Agile Robots unveiled Agile ONE SN at $NVDA’s GTC Berlin, a new humanoid robot built for AI research and development.
CEO Zhaopeng Chen told Reuters the company plans to produce between 1,000 and 2,000 humanoid robots across its ONE S and ONE SN models in 2027, with production expected to begin by the end of 2026.
Here are the key features:
• Size: Stands 1.4 meters (4.6 feet) tall and weighs 40 kg (88 lbs), according to NVIDIA, with 31 degrees of freedom and a 3 kg payload per arm.
• Hands: Two five-fingered WUJI Hand 2 robotic hands, each with 20 degrees of freedom, tactile sensing, and 1kHz force control for precise manipulation.
• AI Computing: Powered by NVIDIA Jetson Thor, delivering 2,070 FP4 teraflops of compute and 128GB of unified memory for onboard reasoning and real-time control.
• Vision and Sensors: Equipped with head and wrist cameras, lidar sensors, and inertial measurement units to perceive and interact with its environment.
• AI Software: Comes preconfigured with NVIDIA Isaac GR00T, providing AI models and tools for robot learning, training, and whole-body control.
• Safety: Planned integration with NVIDIA Halos to support safer robot operation and deployment.
• Development: NVIDIA says the preconfigured platform can reduce setup time for researchers from weeks to days.
• Availability: Expected in early 2027, with Stanford University and ETH Zurich among institutions showing interest.
“The breakthrough for humanoids is usefulness at scale,” CEO Zhaopeng Chen said.
— Wall St Engine (@wallstengine) October 21, 2026
Agile Robots designed this model as a compact research platform. Stanford University and ETH Zurich are among the targeted customers. However, the company has not confirmed any final orders yet.
NVIDIA says its platform brings together hardware, simulation, training, and deployment tools in one place. According to the company, this setup cuts the time researchers need to get started from weeks to just days. Furthermore, NVIDIA product director Spencer Huang said the standardized setup could save researchers two to three months of separately buying and assembling parts.
NVIDIA CEO Jensen Huang also spoke at the event. He said general-purpose robotics is advancing fast and that “the ChatGPT moment for robotics is fast approaching.” He added that the platform gives developers the models, data, and computing systems to bring physical AI to major industries.
Agile Robots was founded in 2018 as a spin-off from the German Aerospace Center. It now employs roughly 2,300 to 3,500 people and targets revenue of about $695 million in 2026. Meanwhile, investors like SoftBank and Xiaomi backed the company in a 2021 funding round that valued it at around $1 billion. Other companies like NEURA Robotics also compete in the growing European humanoid market.
