London-based startup Humanoid has achieved a significant feat in robotics development, moving from concept to a functional HMND 01 Alpha prototype in just seven months. This compressed timeline challenges the traditional 18 to 24-month hardware cycle, leveraging a full-stack NVIDIA platform for simulation, AI, and edge computing.
In the race to build capable humanoid robots, speed of development is as critical as the technology itself. While many projects remain in research labs for years, one ambitious startup is proving that rapid iteration is possible. Humanoid, founded in 2024 by Artem Sokolov, has unveiled its HMND 01 Alpha robots after a remarkably short development sprint. The company, which now employs over 200 engineers across London, Boston, and Vancouver, attributes its velocity to a deep, simulation-first partnership with NVIDIA, creating both wheeled industrial and bipedal research platforms currently in field tests.
“The traditional robotics hardware development cycle is 18 to 24 months. We did it in seven,” the company stated, according to The Robot Report. Central to this acceleration is an integrated software and hardware stack. The HMND 01 Alpha uses the NVIDIA Jetson Thor as its primary edge-computing brain. This powerful platform allows the robot to run large-scale vision-language-action (VLA) models directly on the device, which Humanoid claims has drastically reduced system wiring complexity and improved serviceability. Furthermore, using NVIDIA’s AI infrastructure for training has slashed post-training processing to just several hours, creating a fast loop between data collection and real-world deployment.
READ ALSO: https://modernmechanics24.com/post/orano-selected-doe-900m-uranium-project/
Perhaps the most critical component is Humanoid’s simulation-to-reality pipeline. Built on NVIDIA Isaac Lab and Isaac Sim, this virtual environment is where the robot truly learns. “This virtual training environment allows the team to develop and deploy a new policy from scratch onto physical hardware in approximately 24 hours,” reported The Robot Report. The company also uses these tools for mechanical engineering, evaluating six different leg configurations in simulation to optimize actuator selection and joint strength before any metal was cut. This physics-based optimization was validated in a recent proof-of-concept with automotive giant Schaeffler.
Looking forward, Humanoid is pushing for a fundamental industry shift. “We’re currently working closely with NVIDIA and other partners on a new robotics networking system built on Jetson Thor and the Holoscan Sensor Bridge,” said Jarad Cannon, chief technology officer of Humanoid. The goal is to move away from legacy industrial communication standards toward a modern, software-defined architecture. This approach is gaining traction, with the company already reporting 20,500 pre-orders and three active pilot programs for its wheeled industrial variant. By getting functional systems into the field early, Humanoid aims to gather vast performance data, iterating rapidly on its software to turn a compressed development cycle into a lasting competitive advantage.
WATCH ALSO: https://modernmechanics24.com/post/chinas-humanoid-robot-embroidery-skills/
