New Humanoids Target Heavy Industry

Noble Machines, a Sunnyvale startup emerging from stealth, was founded in 2024 by engineers with experience at SpaceX, Apple, NASA, and Caltech. The company, formerly known as Under Control Robotics, has already deployed its first humanoid, named Moby, to a Fortune Global 500 customer in the heavy industry sector. This rapid deployment, within 18 months of the company's launch, signals a focus on validating its AI and hardware in real-world industrial environments from the outset. Moby is engineered for hazardous and physically demanding "4D" jobs—dull, dirty, dangerous, and declining. It boasts a payload capacity of up to 60 lbs (27 kg), positioning it competitively against other industry humanoids. The robot operates for 5-6 hours per shift powered by a single NVIDIA Jetson Orin and is designed to navigate complex terrains like scaffolding and stairs, a necessity for its target markets in construction, logistics, energy, and manufacturing. Ati Motors' Sherpa Mecha takes a different design route, focusing on functionality over strict human mimicry. It is "humanoid-inspired," featuring two arms and a head, but utilizes a wheeled base for mobility instead of bipedal locomotion. This design choice prioritizes speed, safety, and efficiency in factory environments, avoiding the stability challenges faced by legged robots on uneven industrial floors. The Sherpa Mecha is designed as an adaptable solution for manufacturing, capable of tasks like machine tending, part inspection, and bin handling. It has a payload of 26 lbs at full arm extension and uses 3D LiDAR for navigation and object detection. Ati Motors emphasizes a modular AI approach and seamless integration with existing factory fleet management systems. These robots enter a market where AI is fundamentally shifting capabilities. Instead of task-specific programming, both platforms leverage AI for rapid learning. Noble Machines' Moby uses "whole-body" AI control and can learn new skills in hours through natural language instructions, gestures, and demonstrations. This move toward general-purpose, adaptable robots is a key trend, enabled by foundation models that allow robots to generalize from vast datasets. The push for humanoids in heavy industry is a direct response to labor shortages and the need to improve safety in hazardous roles. Unlike traditional caged robotic arms, these new humanoids are designed to work alongside human crews, fitting into existing workflows and spaces without requiring complete environmental redesigns. However, significant engineering hurdles remain for widespread adoption. Key challenges across the industry include extending battery life beyond the current 2-4 hour operational times to cover full 8-hour shifts and achieving the 95-99% uptime reliability expected in industrial settings. Developing robust AI that can handle unexpected variations in real-world environments is another critical focus area.