2026-01-07



After years of research and development, Boston Dynamics is officially launching its Atlas robot into the market, with Hyundai—the parent company owning 88% of Boston Dynamics—as its first customer.


In October 2025, Boston Dynamics tested a 1.75-meter-tall, 90-kilogram research prototype at a Hyundai factory in Georgia. According to CBS News, the robot autonomously sorted roof racks on an assembly line. The newly unveiled production model is larger and specifically engineered for continuous commercial operations.


Production Model Designed for Warehouses and Factory Floors

The new Atlas stands 1.9 meters tall with a 2.3-meter arm span. It is fully electric, powered by NVIDIA chips, capable of repeatedly lifting 30-kilogram loads in temperatures ranging from -20°C to 40°C. The battery lasts approximately four hours, and Atlas can replace its own battery in just three minutes. All limbs can be swapped out within five minutes.


Hyundai, which holds an 88% stake in Boston Dynamics, will be the first to deploy the robots. A fleet is expected to be delivered in 2026 to the Robot Mass-production Application Center (RMAC). Hyundai is also constructing a new robotics facility designed to manufacture thousands of units annually.


Boston Dynamics anticipates most customers will see a return on investment within two years. CEO Robert Playter stated that the robots are expected to take over heavy, repetitive tasks, though he emphasized that fully replacing human workers may take many more years. Even then, human oversight and maintenance will remain essential.


Atlas Learns via VR Headsets and Motion-Capture Suits

Human involvement remains central to training. Instead of traditional programming, Atlas learns through machine learning. Operators guide the robot using VR headsets or motion-capture suits, while Atlas records their movements. Thousands of digital twins simultaneously train in simulation environments until the system identifies the most efficient methods. Boston Dynamics is collaborating with Google DeepMind and Toyota Research Institute on this advanced R&D initiative.