Agile Robots to build up to 2,000 Nvidia-linked humanoids for research market

Traders Union
All News
Business & Wealth
Companies
Agile Robots humanoids for research

Highlights

Agile Robots plans to produce 1,000 to 2,000 Agile ONE S and ONE SN humanoid robots in 2025 under a new partnership with Nvidia.
The ONE SN model integrates Nvidia's Jetson Thor computer, dexterous hands, cameras, and GR00T AI model for turnkey research and skill-testing applications.
Standardized pre-configuration with Nvidia technology accelerates deployment for customers like Stanford University and ETH Zurich, saving researchers two to three months of integration time.

Production plan and partnership scope

As reported by Reuters, Munich-based Agile Robots plans to produce up to 2,000 petite humanoid robots next year under a partnership with Nvidia announced on Wednesday, CEO Zhaopeng Chen says in an interview.

Production of the Agile ONE S and ONE SN models begins by the end of this year, Chen says. He adds that the company expects to make 1,000 to 2,000 units, with Stanford University and ETH Zurich among potential customers.

The ONE SN model, with the N referring to Nvidia, uses a jointly specified configuration that includes dexterous hands, cameras and Nvidia's Jetson Thor onboard computer. The partnership combines Nvidia computing and its GR00T AI model with Agile's hardware to provide researchers with a ready-to-use platform for collecting data and testing skills intended for later deployment in factories and warehouses.

Research demand and industry positioning

At 1.4 meters tall and weighing about 34 kilograms, the robots are smaller than Agile's full-sized humanoid models, a design that makes them easier and safer for researchers to handle.

Agile, which makes industrial robots, is one of Europe's early contenders in humanoid robotics alongside Germany's NEURA and RobCo, and Britain's Humanoid. Nvidia is establishing an early lead in supplying the processors and software that act as humanoid robots' brains and nervous systems, building partnerships across the emerging sector.

Spencer Huang, product director for robotics at Nvidia, says standardized pre-configuration can save researchers two or three months that might otherwise be spent integrating separately purchased components. He says a system that is already largely assembled can accelerate development work.

In our earlier article on DexForce’s humanoid robot training approach, we explained how the Chinese startup uses synthetic data from virtual simulations to teach robots practical tasks faster and at lower cost than relying mainly on repeated real-world human demonstrations. We also noted that DexForce has attracted major Series B funding and is already seeing commercial deployment, underscoring the industry-wide push to overcome training-data bottlenecks and move humanoids from demos into factories and everyday environments.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.