Why Make a Robot That Looks Like Us, With Phil Zheng-EP 274

Sep 29, 2026

A few years ago, Richtech Robotics would call hospitality groups and offer them robots. They got laughed out of the office. The answer was always the same. We’re a people company.

Now they’re building an 8,000 square foot factory in Las Vegas. Three production lines for robots and sensors, all managed by AI. They’d planned it for 2027, but a new FCC rule made them move it up to this year.

I talked with Richtech’s COO, Phil Zheng, at IMTS. He told me why their humanoid robots have eyes that don’t do anything. Why Chinese citizens might not love robots as much as they seem to. And why even the people building this stuff are surprised how fast AI and robotics are moving. (Blog continues below video)

*************

Link to Graff-Pinkert’s Acquisitions and Sales promotion!

*************

Listen on your favorite podcast app using pod.link.

View the podcast at the bottom of this post or on our YouTube Channel.

Follow us on Social and never miss an update!

Facebook: https://www.facebook.com/swarfcast
Instagram: https://www.instagram.com/swarfcast/
LinkedIn: https://www.linkedin.com/company/todays-machining-world
Twitter: https://twitter.com/tmwswarfblog

*************

Interview Hightlights

The Robot Has to Match the Data

Standing at the Richtech booth, I asked Phil why the robots look like people. The eyes are the giveaway. He admitted straight out that they aren’t meant to do anything.

His answer had three parts, and only one of them was about looks.

The first is the work cell. Most jobs on a production floor are done by a person standing at a station using two hands. If you bring in a robot shaped differently, you have to rebuild the station around it. That costs money and time. A robot with two arms drops into what’s already there. The countertops are already built for us.

The second reason is the one I hadn’t thought about. It’s the training data. Most robot training today comes from teleoperation, where a person puts on goggles and gloves and drives the robot’s arms directly. That’s where the data comes from. So the robot has to have the same shape as the person generating it. You can’t train a robot to work in a shape no human has, because there’s no human in that shape to show it how.

The third reason is the simplest. People like it. They want the head and the eyes. They’ll give the machine a personality, or a name.

Robots That Watch Each Other Work

The humanoid is the thing people take pictures of, but it isn’t the whole pitch. Richtech’s robots are meant to work as a group.

Phil walked me through the table. One robot does assembly and small parts handling, picking and placing, and it comes both mobile and mounted. You don’t want that one hauling things around, so delivery robots move items between stations. Behind them were autonomous pallet jacks for the heavy work.

Tying them together is a fleet management tool and an AI layer that runs all of them. Every robot carries sensors and feeds data back.

Phil’s argument for why that matters was the sharpest thing he said all interview. Right now a person does a step and hands the work to the next person, and nobody has real visibility into what happened in between. With robots, you can see the moment a screw went into a board and how long it took.

The Las Vegas factory is where they prove it out. Sensor assembly on one line, robot assembly on the others, running by the end of this year.

We’re a People Company

The rejection Phil got in the early days wasn’t subtle. He’d call a large hospitality group that couldn’t fill its openings after the pandemic, offer them robots, and get laughed at. We’re a people company.

He says that’s changed. Not because anyone got excited about it, but because people accepted it was going to happen anyway.

I asked about China, since the usual story is that they want robots and we don’t. Phil pushed back on that. He’s ethnically Chinese and he was direct about it. He says regular people in China see robots the same way Americans do, as a threat to their jobs. The difference, he believes, isn’t enthusiasm. It’s that the government and the company owners have the power to install them, and nobody gets asked. In the US we argue about it out loud.

The part he keeps coming back to is the speed. He says it’s moving faster than any technology of this magnitude ever has, and that it’s designed to only go faster.

Question: Would you want a humanoid robot working in your house?