Actuators Are the Real Humanoid Trade

The bet nobody has to make

Every humanoid OEM directory, including this one, lists dozens of companies chasing the same outcome: a robot that can walk into a warehouse, a factory, or eventually a home, and do useful work. Almost none of them are public. Most are pre-revenue, several are pre-prototype, and the ones furthest along are still burning venture money against a market that doesn't really exist yet. Picking the OEM that wins is a real trade, but it's a venture trade: illiquid, binary, and dependent on getting the timing and the company both right.

There's a steadier version of the same bet, and it doesn't require picking a winner at all.

Where the dollars actually sit

Strip a humanoid down to its bill of materials and the actuation stack, the motors, harmonic or cycloidal gearboxes, torque sensors, and the motor-control electronics that drive each joint, is commonly cited as roughly half the total cost, ahead of the compute stack, the battery, and the chassis combined. That's not a detail specific to any one robot. A humanoid needs on the order of thirty to forty actuated joints to do anything a human does, and each one needs this same stack, regardless of whose logo is on the torso.

That's the mechanism the LinkedIn post prompting this piece was pointing at: if the actuation stack is roughly half the cost of every humanoid built by every OEM, the companies making that stack are exposed to the growth of the entire category at once, not to the fortunes of one company inside it. A warehouse-robot OEM and a home-humanoid OEM can have completely different outcomes and still both be customers of the same gearbox or motor-electronics supplier.

Names already in the trade

A few companies with real, established motion-control businesses have humanoid-adjacent exposure already, well before most humanoid OEMs ship a commercial unit. Hyundai Mobis brings decades of automotive-grade motor and actuator manufacturing, sitting inside a conglomerate that also owns Boston Dynamics. Schaeffler's core business is precision bearings and motion systems for automotive and industrial drivetrains, the same mechanical competency a harmonic-drive joint actuator needs. LG Innotek's component manufacturing (motors, sensors, camera modules) already serves automotive and consumer electronics at volume, and volume manufacturing discipline is exactly what actuator suppliers need to hit humanoid-scale unit costs. Smaller, more specialized players, the post named indie.inc among them, are building motor-control silicon specifically for robotics rather than adapting an automotive line.

What these have in common is that they're diversified, revenue-generating businesses today, with humanoid demand as a call option on top rather than the entire thesis. That's a fundamentally different risk profile than an OEM equity stake, public or private.

What to watch, and where this thesis breaks

The honest caveat: "actuators are half the BOM" is a supply-chain argument, not a guarantee any specific supplier captures outsized value from it. Automotive-scale suppliers are used to thin margins and could simply get commoditized the same way in robotics; a vertically-integrated OEM that builds its own actuators in-house (several of the furthest-along humanoid programs already do) removes that OEM as a customer entirely; and a public conglomerate's robotics-adjacent revenue is often a rounding error against its core automotive business, which dilutes the exposure this thesis is actually trying to buy. The trade is real, but it's a supply-chain-concentration bet dressed up as a diversification bet, worth underwriting supplier by supplier, not buying as a basket on the strength of the logic alone.