Noble Machines Opens With Moby: Lifts 60 Lbs, Debuts With Big Customer

Noble Machines just stepped out of stealth with Moby, a humanoid built to haul 60 pounds and run in real-world, outdoor environments.

Noble Machines, a Sunnyvale startup founded by engineers from Apple, SpaceX, NASA, and Caltech, says it delivered its first Moby units to a Fortune Global 500 customer within 18 months of launching from stealth. The company touts a “whole-body” AI control philosophy and rapid, language-based learning as the core of its approach to industrial manipulation. In other words, Moby is supposed to learn tasks not just via grumpy, brittle hand-tuning, but through conversational prompts that steer the robot through new jobs in a factory or yard.

On the hardware side, Moby is described as capable of lifting up to 60 lbs (27 kg) and navigating steep inclines and outdoor terrain. The claim places it in a similar payload bracket as some field-ready humanoids today—lighter than the upper end of Atlas’ strength, heavier than Digit’s payload, and in the neighborhood of a common industrial assist robot when it comes to raw lifting. For context, other humanoids’ publicly cited capabilities include Digit at about 35 lbs, Atlas with a top range of 66–110 lbs, and Figure 3 at around 44 lbs. Noble emphasizes that Moby is designed for hazardous, physically demanding industrial tasks where humans and robots collaborate to keep production moving.

The company’s press notes emphasize an integrated hardware–AI design, aiming to automate dangerous tasks while keeping humans safely in the loop. Demonstration footage suggests a system that can interpret simple instructions, adjust on the fly, and apply the right amount of torque and support to lift or traverse uneven ground. Yet the technology still rides the boundary between “lab-ready” and “field-ready” in the eyes of many engineers. The exiting stealth with a real customer deployment signals a trajectory toward field-scale operation, but it also invites the inevitable questions about reliability, maintenance, and total cost of ownership in day-to-day industrial use.

From a practitioner’s standpoint, two concrete takeaways stand out. First, the language-based learning angle could dramatically shorten task-programming cycles if robust. In practice, that means technician time to teach a robot a new job may shrink from days to hours, provided safety rails and failure modes are well-managed. Second, the “whole-body AI” approach implies cross-functional coordination across locomotion, manipulation, and perception—not just a gripper glued to a robotic arm. That’s potentially powerful for complex jobs like precise part placement on uneven surfaces, but it also raises single-point-failure concerns: if the control stack gets overwhelmed, the entire operation could stall.

The deployment also signals a maturation point for the broader market: a credible, enterprise-facing path from stealth to a Fortune 500 pilot within 18 months is faster than many big-name ambients promised years ago. It suggests growing readiness in systems capable of combining robust on-site autonomy with human collaboration, rather than pure teleoperation or isolated lab demos.

Still, several caveats matter. The official materials do not disclose essential details such as the robot’s power source, runtime, or charging requirements. Nor are DOF counts for Moby or the other players publicly specified in the release, leaving questions about the granularity of the control architecture and how resilient the motion planning will be under real-world grime and weather. As always with a stealth-to-field transition, the proof sits in long-term uptime, serviceability, and the ROI of labor displacement on actual factory floors—areas where many promising prototypes stumble after the demo reel wears off.

If Noble keeps its promises, Moby’s next-generation payload and autonomy improvements could push the company into more aggressive production pilots this year. For now, the field-readiness claim rests on a single Fortune Global 500 deployment in 18 months—a respectable milestone, and one that invites careful, numbers-first scrutiny from R&D teams and investors alike.

Sources

Noble Machines exits stealth with Moby humanoid