Why Agile Robots and Why Now?
Agile Robots is no newcomer. The company was founded in Munich in 2018 by researchers from the German Aerospace Center DLR (Deutsches Zentrum für Luft- und Raumfahrt) — an institution behind the development of robotic technologies for space missions. Since then the company has grown to over 2,500 employees and has raised more than $270 million from investors including SoftBank Vision Fund, electronics maker Xiaomi, and Midas Group.
Today Agile Robots has over 20,000 robotic solutions deployed in industrial facilities worldwide. Under its umbrella it also groups subsidiaries: Franka Robotics (research robotic arms), BÄR Automation (manufacturing automation), idealworks (autonomous mobile robots), and audEERING (emotion recognition from voice).
The partnership with Google DeepMind logically follows the trend of the leading AI lab extending its collaboration into the physical world. DeepMind had previously announced partnerships with other robotics companies — Agile Robots is another step in building a broader physical AI ecosystem.
Gemini Robotics: AI Brain for Industrial Machines
The technical foundation of the collaboration are Gemini Robotics foundation models — base models that Google DeepMind developed specifically for robotics. These models don't just leverage Gemini's language capabilities, but are trained to understand spatial relationships, movement, and physical interaction with objects.
What does this mean in practice? A traditional industrial robot is programmed for precisely defined movements — if a part changes shape or an object shifts a few centimeters, the robot fails. A robot driven by a foundation model can assess the situation, adapt, and respond similarly to a human worker.
The partnership operates on the principle of an "AI flywheel" — a self-reinforcing cycle: robots deployed in industry collect operational data → this data is used to further train models → improved models are returned to the robots → robots become more capable → they collect even more valuable data. The more robots, the smarter the models.
Agile ONE: The Industrial Humanoid Arriving in 2026
The flagship of the product portfolio is Agile ONE — a humanoid robot designed for collaboration with humans in the workplace. Unlike purely industrial arms, Agile ONE can move both arms in a coordinated manner, work in confined spaces, and adapt to different tasks without reprogramming.
The robot enters series production in 2026. Target sectors are those with the highest demand for automation where conventional hardwired machines fall short — situations with variable objects, unstructured environments, or the need for delicate manipulation.
Where Will the Robots Be Deployed?
The partnership will initially focus on four key industrial sectors:
- Electronics manufacturing — assembly of small components, quality control, handling sensitive parts
- Automotive industry — assembly lines, welding, painting, parts logistics
- Data centers — physical server management, component replacement, infrastructure maintenance
- Logistics — package sorting, loading and unloading, warehouse inventory
All these sectors share one thing: a massive labor shortage combined with requirements for precision and reliability that exceed the capabilities of today's generation of robots.
What Do the Leaders Say?
Agile Robots founder and CEO Zhaopeng Chen is convinced we are on the threshold of fundamental change: "The great opportunity lies in autonomous, intelligent manufacturing systems that can transform entire industries."
Carolina Parada, senior director of robotics at Google DeepMind, sees the partnership as a key step in bringing AI to the real world: "This research partnership is an important step toward AI delivering real impact in the physical world."
What This Means for the Global Industry
Agile Robots is headquartered in Munich and has strong ties to the European industrial ecosystem — automotive, mechanical engineering, logistics. These are sectors where manufacturers and suppliers worldwide play a significant role. The combination with Gemini Robotics models could be precisely the moment when industrial robotics stops being just about repetitive movements and becomes truly adaptive.
From the perspective of the EU AI Act, which came into force in 2024, industrial robots with autonomous decision-making are classified as high-risk systems. This means their deployment will require thorough documentation, testing, and transparency — an area where both Agile Robots and Google DeepMind must demonstrate compliance with European rules.
The Race for Physical AI Is Fully Underway
Google DeepMind isn't the only lab betting on physical AI. Figure AI partners with OpenAI, 1X Technologies with Andreessen Horowitz, Tesla Optimus develops a humanoid robot internally. Boston Dynamics, the pioneer of modern robotics, has meanwhile launched commercial deployment of Atlas in Hyundai's manufacturing facilities.
The Agile Robots × Google DeepMind partnership adds a powerful player with a concrete industrial base — 20,000 installations isn't a pilot project, but proven ability to operate in real facilities. Combined with Gemini Robotics models, this could be precisely the moment when industrial robotics stops being about repetitive movements and becomes truly adaptable.
What are Gemini Robotics foundation models and how do they differ from conventional industrial AI?
Gemini Robotics are base models developed by Google DeepMind specifically for physical robots. While conventional industrial AI is programmed for precisely defined movements and fails at the slightest deviation, foundation models understand context, spatial relationships, and can adapt to new situations — similarly to a human. A robot can thus handle a task it was not explicitly programmed for.
When will the Agile ONE robot be available and what will it cost?
Agile ONE enters series production in 2026. A specific price and date for commercial availability have not yet been officially announced. The company is primarily targeting industrial customers in sectors such as automotive, electronics manufacturing, logistics, and data centers.
How will this partnership affect jobs in industry?
Automation of industrial processes inevitably changes the structure of the job market — some routine and physically demanding positions decrease while demand grows for technicians, operators, and AI systems specialists. Agile Robots and Google DeepMind are primarily targeting sectors with an acute labor shortage where robots handle tasks that are dangerous or physically demanding for humans. The long-term employment impact depends on the speed of adoption and the ability of the education system to prepare workers for new roles.