If your company is thinking about buying a humanoid robot, the first question you should ask is not whether the technology is impressive. It’s whether the economics make sense.

Building a robot that can walk on two legs, maintain balance, and manipulate objects with human-like dexterity is one of the hardest problems in robotics. That’s why the engineering coming out of companies like Figure AI , Apptronik , and the humanoid pilots being run by GXO Logistics , are so important to watch.

As someone who has spent much of my career building and commercializing robotic systems, I have enormous respect for the teams tackling this challenge. But impressive engineering and a compelling business case are not the same thing.

Most warehouses, factories, and distribution centers don’t have a “looks like a human” problem. They have a throughput problem, a reliability problem, and a cost problem. When you evaluate humanoid robots through that lens, the picture changes quickly.

The Buyer’s Question Is Economics, Not Anatomy

Operations executives tend to ask the same practical questions when evaluating automation. Will this system move more product per hour? Will it reduce labor costs? How much downtime should we expect? How many technicians will it take to keep the system running? And how long will it take to pay back the investment?

Whether the machine has two arms, two legs, and a head is largely irrelevant. If a robot that looks like a refrigerator can do the job better than a robot that looks like a person, the refrigerator wins every time.

The human body is an extraordinary machine, but it is also incredibly complex. Walking on two legs is inherently unstable. Human hands are mechanical marvels. Coordinating vision, force control, and balance in real time requires a staggering amount of hardware and software.

Replicating those capabilities in a robot is a remarkable technical achievement. It also introduces more actuators, more sensors, more software, and more opportunities for something to fail. Complexity is often the silent killer of automation projects. A system can look flawless in a controlled demo and become far less impressive when exposed to dust, vibration, maintenance delays, and the relentless demands of a 24/7 industrial environment.

Purpose-Built Usually Beats Human-Shaped

Humans are generalists because they have to be. We evolved to operate in environments that were never designed for us. Machines win by doing the opposite. They are purpose-built to eliminate unnecessary motion, unnecessary complexity, and unnecessary variability.

Forklifts don’t need fingers. Conveyors don’t need feet. Palletizers don’t need legs. Robotic arms don’t need bodies.

That’s why the most successful automation systems usually look a little boring. Automated storage and retrieval systems, sortation equipment, robotic palletizers, and autonomous mobile robots are not especially glamorous, but they are repeatable, predictable, and brutally efficient. That is exactly what industrial operators are willing to pay for.

Where Humanoids May Earn Their Keep

None of this means humanoid robots are a bad idea. In fact, I believe they will eventually create meaningful value in certain applications, particularly in environments built entirely around human movement and tools. If a robot needs to climb stairs, open doors, use existing hand tools, and perform a wide variety of low-frequency tasks, a humanoid form factor may prove to be the most practical solution.

But those are still emerging use cases.

The recent humanoid robotics announcements and industrial pilots are important, but they should be viewed as pilots, not proof points. The purpose of a pilot is to answer practical questions about reliability, maintenance requirements, supervision, and total cost of ownership. Until those questions are answered at scale, the business case remains unproven.

The operating principle is simple: the best automation is the one you can afford to operate. That means looking beyond the demo and focusing on what really matters. How often does the system go down? How quickly can it be repaired? How much energy does it consume? What is the true cost per unit moved?

These are not the questions that generate headlines, but they are the questions that determine whether a technology succeeds in the real world.

Humanoid robots are incredible demonstrations of what is technically possible. But real operations don’t pay for what is possible. They pay for what works, consistently, at scale.

The goal is not to mimic humans. The goal is to get the job done better, faster, and cheaper. That’s a very different objective.

I’m bullish on humanoid robotics over the long term. The progress has been remarkable, and there will almost certainly be industrial applications where a human-like form factor offers a meaningful advantage.

But if your company is thinking about buying a humanoid robot today, start with a simple question: Is this robot solving a real operational problem, or are you paying for an impressive demonstration?

In industrial automation, the technologies that win are rarely the ones that attract the most attention. They are the ones that solve real problems and make economic sense. More often than not, they look nothing like us.