For the last two years, the conversation about AI and jobs has focused on white-collar workers. Lawyers. Coders. Analysts. Knowledge workers are wondering if a chatbot is coming for their paycheck.

For us in tech, we see the pressure for software engineers, in marketing, analysis, and anything we do on screens. That’s a real concern.

As I have written, I don’t think it takes all of these jobs, but people will have to learn to use AI to do these jobs in the near future, like starting now.

But while everyone’s been staring at office jobs, something bigger has been happening — and it’s moving a lot faster.

The machines aren’t just coming for knowledge work.

They’re coming for physical work. It’s the robots. They are coming, and they are coming fast.

The people who build things. Move things. Stock things. Fix things. The warehouse worker. The forklift operator. The maintenance tech. The people who were told their jobs were safe because machines couldn’t handle the physical world.

Yeah, well, bad news there, maybe. That assumption is expiring.

And the disruption isn’t coming from ChatGPT.

It’s coming from robots. Smart ones. Humanoid ones.

But wait, there is something else we need to know. Right now, China is building them at a scale that should have every policymaker in Washington losing sleep.

At the end of the day, the rise of robots is only one part of the story. This isn’t just a technology story. It’s a power story. Once again, we find ourselves falling behind in an industry that we basically invented and owned. It’s also an industry and technology that is going to be a huge leadership indicator in the near future.

In 2025, global shipments of humanoid robots jumped to roughly 18,000 units — more than five times the previous year. China accounted for over 85% of those deployments. The United States accounted for about 13%.

One Shanghai startup, AGI Bot, shipped more than 5,000 units by itself — nearly 40% of the global market. Another Chinese company, Unitree, expects to ship between 10,000 and 20,000 humanoids in 2026.

Their starting price: about $13,500.

Tesla’s Optimus, the most visible U.S. effort, isn’t expected to come in under $20,000 even at scale.

When you’re talking about deploying robots across factories and warehouses, that price difference isn’t a rounding error. It’s the ballgame.

But 18,000 units is a pittance. When every household has robots like they have TVs, when robots are in every UPS and FedEx delivery vehicle, when robots are at the bottom of mines, bringing up gold, diamonds and everything else. 18k robots is trivial. We will have millions of robots.

Once the economics make sense, automation stops being a strategy discussion. It becomes inevitable.

China didn’t stumble into this position by the way. It planned for it, it worked for it and it built it.

The sensor supply chains developed for electric vehicles and drones now feed robotics. Government subsidies and policy support have treated humanoid robots as a national priority. Talent programs and manufacturing scale did the rest.

And there’s another advantage that doesn’t get talked about enough. When a Chinese company solves a problem, the solution spreads fast. Knowledge moves across the industry at speed. The whole ecosystem levels up together. It’s sort of like an open source model for robotics. Similar to how the Chinese company behind DeepSeek open-sourced their model (or at least parts of it).

MIT robotics researcher Kim Sang-bae recently put China’s technical position relative to the U.S. at about 98%. Read that again. Ninety-eight percent.

I know you are thinking, well, the U.S. is still ahead, if only by 2%, but while you may take comfort in that, you should know this was a field we invented and owned. The fact that Chinese humanoid robots are roughly on par with ours is a tremendous achievement by the Chinese.  They are already far more widely deployed in industrial robots as part of their manufacturing base.

The Chinese robots themselves are no longer a novelty. A year ago, the headlines were about machines wobbling through dance routines and falling over during public demonstrations. This February, in front of hundreds of millions of viewers, humanoid robots were doing aerial flips, precision choreography and complex coordinated movement. Even Kung Fu demonstrations.

That’s not entertainment. That’s a capability curve moving fast.

The first jobs affected are the obvious ones. Highly repetitive, physically demanding work in structured environments — logistics, warehouse operations, manufacturing, inventory handling.

Walk into a modern fulfillment center today and you’ll already see the future. Fewer people. More machines. Bigger facilities designed around automation.

This isn’t a forecast anymore. It’s deployment.

The harder jobs — the plumber, the HVAC technician, the electrician working in unpredictable environments — will take longer. Robots still struggle in unstructured settings. Reliability matters. Adaptability matters. The AI models controlling these machines still have ground to cover.

But the direction is clear. The physical world is becoming programmable.

What’s less clear is whether the United States plans to compete seriously — or just keep holding roundtables about it, resting on its laurels.

Washington is starting to pay attention. The Commerce Department is convening manufacturers. A Senate proposal would restrict government use of humanoids built by adversarial nations. The Congressional Robotics Caucus has been revived. An executive order is reportedly under consideration.

Those are the right conversations. They’re also years late. We are looking at a robot gap.

American robotics companies are asking for help leveling a playing field shaped by years of Chinese subsidies and a fully integrated domestic supply chain. That’s a reasonable request. These are the kinds of situations where things like tariffs actually make sense. But leveling the field takes time. And China is running up the score.

The part of this that worries me most, though, isn’t the geopolitical competition. It’s who’s missing from the conversation.

Most of the public debate about AI disruption has focused on knowledge workers — people with education, mobility and options. People who can retrain, pivot or move into adjacent roles.

Blue-collar workers have largely been told their jobs were safe. Physical work was supposed to be the last frontier. Machines couldn’t handle real-world complexity. Human labor would always be needed.

That was always temporary. It was always just a matter of time. The machines are catching up faster than most leaders are willing to admit.

China understands the workforce implications. Its government openly talks about humanoid robots as a solution to labor shortages driven by both an aging population and declining birth rates. Robots as a workforce isn’t a dystopian concept there.

It’s policy, because it has to be. They need robotic workers to fill all the physical jobs they need filled, and they don’t have the people to do it. That is different from our situation here in the U.S.

Nevertheless, the United States should be thinking just as clearly — not only about competing globally, but about preparing the people whose livelihoods sit directly in the path of this transition.

This isn’t just the future of technology. It’s the future of who controls work.

Leadership in smart robotics will shape manufacturing, logistics, defense capability and economic leverage for decades. The country that dominates physical automation doesn’t just build products faster. It builds power.

The United States still has enormous advantages — talent, capital, innovation culture, world-class research and entrepreneurial energy. What it doesn’t have right now is urgency and national will and policy.

And in a race where the other side is deploying the future by the tens of thousands of units a year, urgency isn’t a talking point. It’s the difference between leading the next industrial era… Or watching it happen somewhere else.