Automakers and Humanoid Robots: Early Factory Pilots, Real Risks, and Too Much Hype

The surrender story is too simple

The automaker story is not a clean split between bold robotics companies and asleep-at-the-wheel carmakers. It is messier and more useful than that. Some automakers are quiet. Others are already testing humanoid robots or backing robotics through partnerships.

That matters because factories are one of the first plausible homes for humanoids. They have structured tasks, safety procedures, repeatable workflows, and expensive downtime. They also expose the hard part quickly: a robot has to do useful work over and over, not just look impressive in a demo.

What is actually happening

• BMW has tested Figure robots in factory work, and Figure has published its own BMW deployment material.
• Mercedes has worked with Apptronik around Apollo-style logistics and manufacturing tasks.
• Hyundai owns Boston Dynamics, whose Atlas program is explicitly aimed at industrial use and customer testing.
• Tesla continues to position Optimus as part of its AI and robotics work, although external availability and repeatable factory ROI need proof.

Why automakers care

Cars made the modern factory, and factories are where industrial robots already have the most history. IFR still shows automotive as one of the central robot-using sectors. Humanoids appeal because they might work around existing tools, carts, doors, shelves, and human-scale layouts.

The value is not magic labor replacement. The value, if it arrives, is flexibility: moving parts, inspecting, carrying, loading, unloading, and doing ergonomic work where a fixed industrial robot is too rigid or too expensive to integrate.

China and EV-linked robotics

Chinese EV and robotics companies are moving quickly, and XPeng's IRON, Unitree's lower-cost platforms, UBTECH, AgiBot, and others deserve close attention. But the article should not treat every Chinese automaker as having a proven humanoid strategy unless the company has actually published one.

What remains unproven

• Can the robot complete a useful factory task for a full shift with low supervision?
• Can it operate safely around people without slowing the process down?
• Does it reduce cost after support, integration, downtime, insurance, and teleoperation are included?
• Can vendors service fleets quickly enough for production environments?
• Does the automaker keep useful data and process knowledge, or only buy a flashy pilot?

The factory lesson

Automakers have not silently surrendered. They are in the early part of an uneven industrial robotics race. The winners will not be the companies with the loudest demo. They will be the ones that turn narrow pilots into safe, boring, repeatable work.

Compare on RoboZaps: Atlas. Full specifications and a quote route for each, alongside every humanoid robot we carry.