Humanoid.ai
HMND 01
In summary, HMND 01 is a highly capable humanoid robot engineered for automation, precision, and adaptability. With its advanced movement, AI-driven perception, and industrial-grade functionality, it has the potential to transform various industries by streamlining operations and supporting human workers.
Description
The HMND 01 humanoid robot, developed by UK-based Humanoid.ai (also known as SKL Robotics Ltd.), represents a groundbreaking advancement in commercially scalable humanoid robotics, launched in alpha prototypes throughout 2025. This modular platform comes in two primary variants: HMND 01 Alpha Wheeled and HMND 01 Alpha Bipedal, each optimized for industrial automation, logistics, manufacturing, and potential future service applications. The wheeled variant stands at 220 cm tall, weighs 300 kg, and achieves a maximum speed of 2 m/s (7.2 km/h) with a 4-hour runtime per charge, while the bipedal version measures 179 cm, 90 kg, 1.5 m/s (5.4 km/h), and 3-hour runtime. Both share a 15 kg payload capacity and 29 degrees of freedom (excluding end-effectors), enabling human-level or superior manipulation speeds.
Architecturally, HMND 01 integrates robust, modular hardware with cutting-edge AI. The design allows interchangeable upper/lower bodies, end-effectors (12-DOF 5-finger hands or 1-DOF parallel grippers), and protective garments for environmental adaptation and safety. Sensors include 360° RGB cameras, dual depth sensors in the head, wrist RGB cameras, 6D force/torque (F/T) sensors in torso/wrists/end-effectors, haptic feedback, a 6-mic array, speakers, and a head display with LEDs for multimodal human-robot interaction. Processing is powered by NVIDIA hardware (likely Jetson Orin AGX equivalents) paired with Intel i9 CPUs, supporting Ethernet/Wi-Fi connectivity.
The AI stack is centered on Humanoid's proprietary VLM (Vision-Language Model) and VLA-based (Vision-Language-Action) KinetIQ framework, enabling end-to-end reasoning, perception, and action. Locomotion employs reinforcement learning (RL)-trained Whole-Body Controllers for bipedal feats like straight/curved walking, sidestepping, backward motion, squatting, hopping, running, and omnidirectional push recovery. Wheeled mobility offers omnidirectional navigation at higher speeds. Software emphasizes low-latency, model-based, and learning-based controls for precise loco-manipulation in confined spaces, with support for autonomous/teleoperated modes and fleet scalability.
Real-world deployment history began with proofs-of-concept (POCs) in late 2025. In October, partnered with Schaeffler for bin picking metallic bearing rings in a near-production environment at Erlangen, demonstrating cluttered bin retrieval, multi-station mobility, and rapid VLA skill adaptation. November's SAP pilot targeted intelligent kitting/tote handling for automotive seat assembly with Martur Fompak. Demos include holiday tasks like gift-wrapping (December 2025) and CES 2026 live shows of bin picking. Plans include retail pilots and mass commercialization, addressing labor shortages via low TCO and human-safe collaboration. With rapid iteration—wheeled in 7 months, bipedal walking in 48 hours—HMND 01 positions Humanoid.ai at the forefront of Europe's humanoid race, blending industrial reliability with adaptive intelligence for transformative automation.
Key Features
Modular Design
Interchangeable upper/lower bodies, end-effectors (12-DOF hands or grippers), and garments for customization across industrial scenarios, reducing costs and enhancing adaptability.
Advanced Sensing Suite
360° RGB cameras, depth sensors, 6D F/T sensors, haptic feedback, 6-mic array, and head display for full environmental perception and human interaction.
VLA-Powered AI (KinetIQ Framework)
Vision-Language-Action models with end-to-end reasoning, RL-trained whole-body control for autonomous task execution, rapid skill acquisition, and human-level manipulation.
High-Performance Locomotion
Bipedal: walking, hopping, push recovery up to 1.5 m/s; Wheeled: omnidirectional up to 2 m/s, enabling navigation in confined spaces.
Industrial Reliability
Low TCO design with robust hardware/software for fleet deployment, teleoperation support, and applications in logistics, manufacturing, and warehousing.
Specifications
| Availability | Prototype |
|---|---|
| Nationality | UK |
| Website | https://thehumanoid.ai/ |
| Degrees Of Freedom, Overall | 41 |
| Height [Cm] | 175 |
| Manipulation Performance | 2 |
| Navigation Performance | 2 |
| Strength [Kg] | 15 |
| Weight [Kg] | 70 |
| Runtime Pr Charge (Hours) | 4 |
| H.G Skill Score | 4 |
| Verified | Not verified |
| Manufacturer | Humanoid.ai |
| Variants | alpha_bipedal: {'height': '179 cm', 'weight': '90 kg', 'dof_overall': '29 (excl. end-effectors)', 'dof_hands': '12 per hand (5-finger)', 'max_speed': '1.5 m/s (5.4 km/h)', 'payload': '15 kg', 'runtime': '3 hours', 'sensors': '360° RGB cameras, 2x depth sensors, 6D F/T sensors (torso/wrist/end-effector), haptic feedback, 6-mic array, speakers, head display/LEDs', 'processors': 'NVIDIA (Jetson Orin AGX equiv.), Intel i9', 'connectivity': 'Ethernet, Wi-Fi', 'os': 'VLM/VLA-based KinetIQ framework', 'locomotion': 'RL-trained whole-body control (walking, sidestep, hop, push recovery)'}, alpha_wheeled: {'height': '220 cm', 'weight': '300 kg', 'dof_overall': '29 (excl. end-effectors)', 'dof_hands': '12 per hand (5-finger) or 1-DOF gripper', 'max_speed': '2 m/s (7.2 km/h)', 'payload': '15 kg', 'runtime': '4 hours', 'sensors': '360° RGB cameras + depth, wrist RGB, 6D F/T, haptic', 'processors': 'NVIDIA-powered VLA intelligence', 'base': 'Omnidirectional wheeled'} |
| Common | materials: Industrial-grade with interchangeable garments, thermal_management: N/A (industrial robustness emphasized), battery: N/A kWh specified; supports 3-4h runtime, motor_tech: High-torque, high-speed actuators, ip_rating: N/A, llm_integration: Yes via KinetIQ |
Curated Videos
