S1
The S1 is a Humanoid humanoid robot developed by Galbot. The S1 is unknown.
Primary use case: General-purpose embodied AI robot for autonomous manipulation in physical-world tasks
Specifications
Physical
Locomotion
unknown
Manipulation
Hand Type
unknown
Power & Autonomy
Autonomy Level
unknown
AI & Compute
AI Capabilities
vision, planning, manipulation
Commercial
Availability
unknown
Primary Use Case
General-purpose embodied AI robot for autonomous manipulation in physical-world tasks
Related news
Agility is public: the short list of humanoid robot stocks, and who's next
The engineer who taught 1X's NEO to walk just raised $30 million to give robots a sense of touch
Hong Kong bets HK$10 billion on becoming the world's humanoid robot testing ground
Google one-shots humanoid training, Tesla gives robots skin, Faraday launches $2,490 humanoids, and more!
Galbot's humanoid already runs 200 stores in China. It just opened 3 more in Hong Kong.
Skild AI trains robots to learn physical skills through self-play
Skild AI has demonstrated a self-play training approach that enables robots to develop physical skills without human demonstrations, using its S1 robotics foundation model. The model learned soccer through more than 140 years of simulated play, developing behaviors like dribbling, tackling, and recovering from falls, before transferring the policy to a physical humanoid robot. The approach is being extended to multi-agent scenarios and everyday robotics tasks beyond sports.
Skild AI trains robots to learn physical skills through self-play
Skild AI has demonstrated a self-play training approach using its S1 robotics foundation model that enables robots to develop physical skills without human demonstrations, tested by training a robot to play soccer through more than 140 years of simulated play in Nvidia Isaac Sim. Given only the high-level objective of scoring a goal, the policy competed against progressively stronger versions of itself, with complex behaviors such as dribbling, tackling, shielding, and recovery from falls emerging naturally. The resulting policy was successfully transferred to a physical humanoid robot, and Skild sees this as a path toward physical AGI applicable to factories, construction sites, and homes.
Galbot’s PASSAGE Helps Humanoids Step Over, Duck Under and Squeeze Past Obstacles
Galbot has highlighted PASSAGE, a research framework enabling a humanoid robot to choose how to traverse cluttered environments by stepping over low obstacles, ducking under overhead barriers, or turning sideways through narrow gaps. The system pairs a motion planner with a whole-body controller trained on 100 hours of human VR motion capture data, reaching an 89.67% destination success rate in 750 simulated trials and successfully completing all 50 physical trials on a Unitree G1 across unseen layouts.