Humanoid Robot Safety Debate Puts Actuator Failure on Buyer Checklists

Industry Update

Drives&Controls has published an article by Dr David Brandt, vice-president of R&D and chief technology officer at Teradyne Robotics, asking how humanoid robots should respond when a component fails in a real industrial environment. The article was published on September 7, 2026, and frames the issue around safety, trust, and practical factory deployment rather than demonstrations alone.

The article notes that recent robotics safety work, including Nvidia's Halos for Robotics, puts heavy emphasis on perception, external sensing, and collision avoidance. Brandt argues that this is useful, but that functional safety also needs to ask what happens when the robot is no longer fully in control of itself.

For dynamically balancing humanoids, the article highlights a specific concern: a legged robot may fall if an actuator, sensor, battery, or balance-control function fails mid-task. It contrasts that risk with conventional industrial robots, cobots, and mobile platforms that can often move toward a simpler safe state because they are designed around a stable base or a more predictable motion envelope.

The article also references standards context. It says industrial robots are typically designed around single-fault tolerance principles under ISO 10218-1, while ISO/CD 25785-1, a standard for dynamically stable mobile robots, is still moving through committee review.

Buyer Considerations

For automation buyers, the immediate takeaway is not that humanoids are unusable. It is that robot selection should include a documented failure-response review before buyers compare payload, reach, software features, or vision capability. Actuator redundancy, brake behavior, battery fault handling, and controlled-stop strategy can be as important as perception performance.

Buyers comparing humanoids with cobots, AMRs, gantry systems, or other purpose-built automation should ask vendors to show stability data, single-fault assumptions, and the applicable safety standard path. This is especially important in workflows where workers share space with moving equipment, or where a falling machine could create a hazard even after collision detection has worked correctly.

For motor and drive sourcing teams, the discussion also points back to component-level requirements. Servo actuators, encoders, brakes, gearboxes, battery management, and drive diagnostics all contribute to a credible safety case. Procurement teams should avoid treating these parts as interchangeable commodities when the machine depends on coordinated motion to remain stable.

Frequently Asked Questions

What changed for robot buyers?

The article adds a fresh safety lens to humanoid robot evaluation: buyers should ask how the system behaves after a component fault, not only whether it can see and avoid people during normal operation.

Does this affect motor and drive selection?

Yes. If a robot needs continuous motion control to stay stable, actuators, encoders, brakes, drives, and power systems become part of the safety discussion, not only performance specifications.

Are humanoid robots ruled out for factories?

No. The article says the safety case is unfinished and should be examined carefully. Buyers should compare humanoids with cobots, AMRs, and other purpose-built machines based on the task and the verified safe-state behavior.

What should buyers ask suppliers?

Ask for failure-mode analysis, stability data, single-fault assumptions, applicable safety standards, controlled-stop behavior, and how the robot handles actuator, sensor, and battery faults during a task.