Operator's Read: The Robots Aren't Ready. The Need Is.

What the factory photos don't show you.

Frank J. Lazowski III  |  Operis Global  |  July 2026

You have seen the videos. A humanoid robot walking a factory floor, sorting parts, carrying a tote, standing shoulder to shoulder with people on the line. The message is obvious. The future has arrived, and it walks on two legs.

Here's what the photo doesn't show.

The fence.

And here is why any of this matters in the first place. Roughly half a million manufacturing jobs sit open in the United States right now, and Deloitte and the Manufacturing Institute project the sector will need 3.8 million workers through 2033, with as many as 1.9 million of those jobs going unfilled if nothing changes. That is the real story. Not that the robots are impressive, but that the work is there and the people are not.

I spend my time on factory floors, and I was at Automate this year. The booths were full of humanoids. Very few are ready for real work, and in almost every live deployment the robot is either walled off from people or slowed to a crawl. To be clear, I believe humanoids are coming. I just do not believe they are arriving on the timelines shown in most marketing videos.

Here's my operator's read.

The money finally showed up.

For years the oxygen went to AI. Large language models got the headlines and the valuations while robotics did the unglamorous work of moving atoms instead of tokens. That has flipped. Humanoid startups raised about $6.1 billion in 2025, more than triple the year before. Figure is valued at $39 billion. Goldman Sachs now sees a $38 billion market by 2035, after revising its own forecast up sixfold.

But capital is not capability. Money builds robots. It does not certify them, staff them, or make them safe next to a human being. We saw this movie in reshoring: $1.7 trillion announced, and the binding constraint was people, not dollars. Same lesson here.

China is leading. Again.

While the West is running high-profile pilots, China is building units. Morgan Stanley expects China to ship around 50,000 humanoid units this year, nearly double its earlier forecast. Chinese makers already account for the large majority of the world's production, backed by a hardware supply chain built through the EV boom, from actuators to batteries to sensors. Tesla's Optimus is not a volume product yet. The US leads on AI. Europe holds premium and medical niches. But on the metric that matters most to manufacturers, building and deploying reliable units at cost, China is ahead.

Look at what they actually do.

Watch the pilots. The robot moves a tote. Loads a machine. Carries material. That is valuable work, and it is worth automating. But do not confuse it with skilled manufacturing. A machinist. A toolmaker. A maintenance technician. Those jobs are not disappearing.

And as an industrial engineer, I will tell you the time study is not flattering. A humanoid running a welding torch works at roughly one-fifth the speed of a fixed arm. On a high-volume line, that is not a competitive option. It is a science project. Cycle time is money.

Safety hasn't caught up.

This is the part the videos skip. The standards have not caught up with the machines, so real deployments fall back on two crude tools: wall the robot off, or slow it to a crawl. Both work. Both cap the value. A robot behind a fence is often just an expensive, awkwardly shaped industrial robot, and we already have better, cheaper ones of those.

So what does an operator do?

I spent years in industrial automation. I learned not to ask whether a technology is exciting. I ask whether it survives first shift on Monday morning. For most humanoid use cases today, the honest answer is not yet.

So do the basics. Fix the process first, then deploy what is already proven and already safe: cobots, autonomous mobile robots, and well-designed fixed automation, now married to AI that makes those systems smarter than they were two years ago. Those solutions exist today, they are safe today, and they pay back today.

And remember the cobot. Nearly two decades after it arrived, it is still only about 14 percent of industrial robot installations, because it took years to learn where it fit. Humanoids are earlier and more complex. This will take time.

The future is coming.

The labor math guarantees it.

But it will arrive on the factory floor before the living room, in China before the West, and behind a fence before it walks freely beside your people.

Build for the future. Operate in the present. Invest where the ROI exists today. Prepare for where the capability will exist tomorrow.

Separate demonstrations from deployments. Separate headlines from operating reality. The winners will not be the companies that bought the first humanoid. They will be the ones that knew exactly when the technology was ready.

Operis Principle No. 5: Capital is not capability. The demo is not the deployment.

Clarity. Focus. Execution. Value.

Frank Lazowski is the founder and Managing Director of Operis Global, an operating-advisory practice for industrial, manufacturing, automation, and PE-backed companies. Over more than 35 years, he has built and led global manufacturing, sourcing, supply chain, and automation operations across four continents.