Self-developed native world model with key breakthroughs in physical cognition and intelligent decision-making
Flagship dexterous hand with 20 DoF and 44 high-resolution 3D tactile sensors
Enterprise-grade open-source VLA engineering base, advancing scaled robot deployment
Full-stack self-developed robot system; fully in-house software and hardware, based on the integrated 3D world model + reinforcement learning design, purely simulation-trained; grasps 100+ unseen objects nonstop for 60 minutes
The world's first embodied foundation model on the World Unified Model (WUM) architecture: vision, language, touch, action and physical prediction jointly trained in a single network, eliminating inter-module boundaries and data-transfer losses
Second-generation all-round household dual-arm robot, 21 DoF, 24 sensors, full-reach mobility and fine force control; covers childcare, light housework, pet interaction, and elder companionship
The world's first brain-inspired VLA model and the only one combining three bio-motion capabilities: active perception, fault self-recovery, and temporal memory; open-sourced on the AlphaBrain Platform in Jun 2026
The first vertically integrated large-scale humanoid robot factory in the US, in Hayward, CA; reached full production on Apr 30, 2026
1X TechnologiesHumanoidNinebot mobile base plus dual robotic arms, coordinating manipulation and mobility to close the loop from back kitchen to front-of-house service
Autonomous-driving foundation model with a native multimodal MoE Transformer architecture; five innovations: 3D spatial understanding, multimodal reasoning, unified action generation, closed-loop reinforcement learning, and hardware-software co-design
Spot 等机器人搭载 FFM 实现「charted territories 之外」的自主巡检，识别坑洞、坠落风险、PPE 合规等
Field AISoftware / Data / ModelsGeneral embodied foundation model; the world's first general embodied foundation model that 'can actually work'; set the Guinness World Record for most sub-millimeter wire-harness assemblies in one hour
「robots-to-goods」全自主履约系统（AI 驱动），2026 年 LogiMAT/MODEX 展出，获 2026 RBR50 奖
Consumer bipedal robot; the world's first 'mind-native' small bipedal robot, 95 cm tall and 20 kg
Newest small quadruped, 5 m/s, can climb 50 cm platforms
The industry's first interactive world action model; the DAWN architecture compresses visual features into 16 core semantic tokens, cutting inference latency to 331.3 ms; video-generation prediction + VLM macro causal constraints
A VLA model for the clothes-folding vertical scenario
Embodied Agentic OS; agent + physical intelligence dual engines, breaking through 'lifelong learning' and 'cross-scenario generalization'
与 Lucid、Uber 合作的生产级自动驾驶出租车，基于 Lucid Gravity SUV+Nuro Driver，CES 2026 发布，2026 年湾区商用
