Summary:A London-based robotics startup has rapidly advanced its HMND 01 humanoid system to Alpha prototype stage, leveraging NVIDIA’s edge AI and simulation stack to accelerate development and real-world deployment.

A robotics startup has progressed its HMND 01 system from concept to Alpha prototype in just seven months, significantly shortening the traditional 18–24 month hardware development cycle. The system now includes both wheeled industrial and bipedal research robots, with early field testing already underway.

Humanoid follows a simulation-first approach using NVIDIA Isaac Lab and Isaac Sim. | Credit: Humanoid

The HMND 01 Alpha platform is built around an NVIDIA edge computing architecture, enabling on-device execution of large-scale vision-language-action models. This reduces system complexity, improves wiring efficiency, and allows faster iteration cycles, with model updates and training feedback loops compressed to just hours.

A key feature of the development process is its simulation-first workflow powered by NVIDIA Isaac Sim and Isaac Lab. Engineers use digital twins and reinforcement learning to train robotic behaviors in virtual environments before deploying them to physical robots, enabling hardware and software validation cycles as short as 24 hours.

Simulation tools are also used for mechanical design optimization, including leg structure evaluation, actuator selection, and sensor placement. This approach reduces real-world testing risks while improving stability, perception coverage, and operational reliability in industrial environments.

Founded in 2024 with offices across London, Boston, and Vancouver, the startup is positioning its wheeled robots for near-term industrial deployment while developing bipedal models for broader applications. The HMND 01 platform is reported to have attracted over 20,000 pre-orders and is currently undergoing multiple pilot programs focused on logistics and industrial automation.

Comparing HMND 01 Alpha wheeled and bipedal versions