Booster T1

The Booster T1 is a bipedal humanoid robot from Shenzhen-based Booster Robotics, designed for industrial and logistics applications. The T1 features robust bipedal locomotion with a focus on stability in real manufacturing environments, dexterous two-arm manipulation, and an open SDK for developer integration.

Primary use case: Industrial operations, manufacturing assistance, logistics

Related news

New IDC data: humanoid robot shipments up 432% year over year, jobs for robots, and much more ...

D-Robotics just raised $400M to build the brain inside your next robot

Tesla ramping Optimus production, D-Robotics wants to be the NVIDIA of China, and 13,361 people ordered a home robot

From around the web

Booster Gym: An End-to-End Reinforcement Learning Framework for Humanoid Robot Locomotion

Booster Gym is an open-source, end-to-end reinforcement learning framework for humanoid locomotion developed by Tsinghua University and Booster Robotics Technology Co., Ltd, covering training, domain randomization, reward design, parallel-mechanism conversion, and deployment. It was validated on the Booster T1 humanoid robot, where policies trained with PPO in Isaac Gym transferred zero-shot to hardware, enabling omnidirectional walking, disturbance recovery, and locomotion on varied terrains and slopes up to 10 degrees without explicit terrain perception.

alphaxiv.org•Oct 5, 2026

New 56% Benchmark in Figure Triggers Critical Generalization Verification Challenge for All Embodied Agents

Figure AI released Helix 2.5, achieving a 56% zero-shot generalization success rate across 30 unseen homes—a leap from 9%—using its Index dataset of 16 million videos and a new scaling approach. Domestic competitors like Unitree, Agibot, and Galaxy Universal are responding with their own data scaling curves, third-party rankings, and specific application deepens, with Wang Xingxing proposing an 80% threshold as the industry explosion point within 2-10 years.

eu.36kr.com•Sep 20, 2026

Booster Robotics Powers Science Robotics Research Article on Vision-Driven Humanoid Soccer Skills

A research article published in Science Robotics presents a unified reinforcement learning framework enabling humanoid robots to acquire reactive soccer skills through direct integration of visual perception and motion control, trained entirely in simulation and deployed zero-shot on Booster Robotics' T1 platform. The framework demonstrates a 46% reduction in ball position estimation error, 64% reduction in time-to-kick, and approximately 90% kicking success rate in frontfield positions.

theglobeandmail.com•Aug 21, 2026