Atom
A full-size bipedal humanoid robot by Dobot Robotics standing 1.53 m tall and weighing 62 kg with 28 degrees of freedom and ±0.05 mm arm precision, powered by a 1500 TOPS edge AI compute module running Dobot's Robot Operator Model-1. Commercially available from approximately $27,500 and in mass production since 2025.
Primary use case: Electronics manufacturing, precision assembly, logistics, and research
Height
153 cm
Weight
62 kg
DOF
28
Price
$28K
Specifications
Physical
Height
153 cm
Weight
62 kg
Degrees of Freedom
28
Locomotion
biped
Commercial
Price
$28K
Availability
available
Primary Use Case
Electronics manufacturing, precision assembly, logistics, and research
Related news
General Robotics CEO: we're wasting billions on physical AI chasing the wrong layer
Clone shows off an amazing robot hand it's already retiring. The new hand might be totally different.
Inside China's robot schools: 70 training grounds where humanoids drill 45 skills up to 600 times a day
Hello construction: this semi-humanoid robot can use tools and build houses
Figure's billion-dollar bet, and the company that won the robot Olympics
Atom and NSK Announce Strategic Partnership to Help Develop the Humanoid Robot Industry in Japan
Atom and NSK have announced a strategic partnership through a memorandum of understanding to jointly develop and test humanoid robot components and AI systems in Japan. The collaboration will focus on testing NSK-developed actuators on Atom humanoid robots and using NSK manufacturing sites to collect physical AI data for model improvements and advanced manufacturing applications.
AdaPT Bridges Sim-to-Real to Teach Humanoids the Tennis Styles of Federer, Nadal, and Djokovic
Researchers from Shanghai AI Lab, Dobot, SJTU, and Noitom introduce AdaPT, an adaptive planning and tracking framework that extracts stylized tennis rallying and serving from broadcast footage.
Two Days After Releasing HiPHI, Noitom Robotics and Collaborators Demonstrate Humanoid Robots Playing Professional-Style Tennis
Noitom Robotics and collaborators unveiled AdaPT, a system that teaches humanoid robots to play professional tennis by pre-training on the HiPHI-series motion data. The system was validated on Unitree G1 and Dobot Atom robots, demonstrating rally and serve skills in the styles of Rafael Nadal, Roger Federer, and Novak Djokovic. The demonstration was held two days after the public release of the HiPHI dataset at the World Robot Conference, ahead of RO-MAN 2026.