Are humanoid robots going to decide which countries get rich and which fall behind? Probably yes.
In this TechFirst, I talk with Dr. Robert Ambrose, former head of one of NASA’s first humanoid robot teams and now chairman of Robotics and Artificial Intelligence at Alliant. We dig into the future of humanoids, how fast they are really advancing, and what it means if China wins the humanoid race before the United States and other western nations.
We start with NASA’s early humanoid work, including telepresence robots on the space station that people could literally “step into” with VR in the 1990s. Then we zoom out to what counts as a robot, why bipedal mobility matters so much, how humanoids will move from factories into homes, and why the critical photo of the robot revolution might be taken in Beijing instead of Times Square.
Along the way, Ambrose shares how US policy once helped avoid losing robotics leadership to Japan, why the National Robotics Initiative mattered, what the drone war in Ukraine is doing to autonomy, and how small and medium businesses can survive and thrive in a humanoid and AI agent world.
Check it out here and don’t forget to subscribe to my YouTube channel:
Summary of this episode: humanoid robots … USA vs China
Humanoid robots are no longer sci fi, but they are also not quite here yet. They are somewhere in between, and where they go next is likely to decide which countries get rich and which fall behind.
Talking with Dr. Robert Ambrose, former head of one of NASA’s first humanoid robot teams and now chairman of Robotics and AI at Alliant, a few themes really stood out.
1. Humanoids are built for a world made by humans
First, humanoids exist because the world is built for humans. NASA’s early robots were not designed to be cool space androids. They were designed to use the tools and interfaces already on the shuttle and the ISS: latches, levers, power tools, handles, switches. Redesigning space infrastructure for simple robots was impossible, so NASA built robots that could live in a human environment. That logic applies just as strongly to factories, warehouses, and eventually homes.
Ambrose’s early work at NASA included a powerful “stepping into the robot” moment. In the 1990s his team had a teleoperated humanoid upper body with a VR headset and data glove. When operators put the system on, moved their hands, and saw robot hands move in sync in VR, they felt physically present in the machine. People would drop an object and instinctively move their feet to get out of the way, even though the robot did not have toes and they were across the room. That sense of embodiment showed how strong telepresence can be and hinted at what remote work through robots might look like.
2. Hands are the critical innovation in humanoid robotics
Second, humanoids are mostly about hands and mobility.
In Ambrose’s view, a humanoid is essentially a work system (upper body and hands) attached to a mobility system (legs or wheels) whose job is to move those hands into the right place. Biped locomotion is not a party trick. It is what lets robots get in and out of cars, squeeze between parked vehicles, climb around complex environments, and operate in spaces no Roomba or tank-style bot will ever handle without redesigning everything around them.
At the same time, Ambrose is very clear about “design tension” between purpose built machines and general purpose humanoids. For some jobs, a single purpose machine is obviously better: a Roomba versus a humanoid with a vacuum cleaner. He expects we will always have a large universe of special purpose machines that many people will not even think of as robots at all. In fact, he has a very big tent definition of robots: elevators, advanced soda machines, Roombas, and humanoids all qualify as long as there is motion, sensors, and computer control. Over time, he thinks many of those “robots” will just fade into the background as machines, while the truly general purpose systems remain our archetypal robots.
3. Innovation is not even or smooth … it’s “lurchy”
Third, robotics moves in “lurches,” not smooth curves. Industrial robot arms for welding and painting in the 1980s were one lurch. Mars rovers and then DARPA’s self driving car challenges in the 2000s were another. Cheap, self stabilizing drones that anyone can fly were another. Each jump was driven by a combination of government investment, difficult constraints, and real world necessity.
Bipedal humanoid mobility is the latest lurch. Over roughly a decade it has gone from blooper reel material to marathon distance achievements. The hands are now mobile in a human world, which unlocks entirely new categories of work.
4. Humanoids will eventually be very affordable for most
Fourth, humanoids follow the same economics as cars and PCs. Early automobiles were expensive, unreliable, and niche. Once price dropped and reliability went up, they transformed work, cities, and entire nations. The same thing happened with personal computers: once business use drove manufacturing scale, consumer machines exploded.
Ambrose expects humanoids to follow the same curve. Early units will be “trucks” for industry: high value, factory focused, bought by companies. That industrial demand will drive volume, push prices down, and improve reliability. Eventually a consumer version appears: the Rosie stage, where a family can buy a humanoid because it delivers more value than its price.
He uses a simple rule of thumb: when a humanoid can reliably create 10,000 dollars worth of value and costs 9,000 dollars, people will buy them. That might start in factories and warehouses, but it does not end there. At some point, average families will purchase general purpose robots that do cleaning, light security, elder care, child monitoring, pet watching, and basic household tasks. Ambrose also makes the important point that there is no fundamental reason a humanoid should cost more than a car once you get to scale. Similar amounts of material, similar numbers of motors and actuators, similar complexity. The only missing ingredient right now is volume and maturity.
5. The geopolitical implications are huge
Fifth, the geopolitics are huge. The 20th century was, in many ways, the car century, and that helped make it the American century. The iconic “before and after” photos of Times Square filled with horses and then cars, less than a decade apart, captured a massive shift in economic power, infrastructure, and industrial ecosystems that the US owned.
Ambrose’s concern: what if the equivalent “before and after” photos of the humanoid transition are taken in Beijing instead of Times Square. China already has deep manufacturing ecosystems, co located suppliers, and strong state support. It is investing heavily in humanoids to address both external industrial goals and internal demographic challenges, like a shrinking workforce and aging population.
If China gets to cheap, reliable humanoids first at scale, the country with the existing manufacturing backbone and integrated supply chains could lock in a new century of industrial advantage. Robots are almost free labor when amortized over long periods and high utilization. Combine that with already concentrated manufacturing, and the gap can grow, not shrink.
6. Government investment and government policy can be huge levers for growth
Sixth, policy and investment choices matter. Twenty years ago, US robotics leaders assumed they had already “lost” robotics to Japan. That spurred Ambrose and others to work with the White House Office of Science and Technology Policy and multiple agencies on the National Robotics Initiative. It funded university research, built a pipeline from robotics-obsessed K-12 kids to university programs, and helped spawn a wave of robotics startups in the US. That may be the main reason the US is still competitive in robotics today.
Government investments like DARPA challenges, Mars rovers that must drive autonomously on another planet, and even grim realities like the drone war in Ukraine all shape capability. The same will be true for humanoids. If the US does nothing, Ambrose thinks China will dominate, with Japan possibly overtaking the US as well. If the US acts, there is still time to stay in the lead or at least remain a true peer.
What he argues for is not a command and control industrial strategy, but targeted support for talent and deployment: education, R&D, and incentives for companies to adopt and manufacture robots domestically. Innovation and entrepreneurship are America’s unfair advantage, especially when combined with chaotic, bottom up experimentation that centralized systems have trouble matching. But that only pays off if innovations are manufactured and scaled at home instead of being shipped overseas.
7. Robots and people will work together
Seventh, the future of work is not robot overlords and mass unemployment. Ambrose is dismissive of both robot apocalypse narratives and “robots will take all our jobs” fatalism. We are in a prolonged period of very low unemployment, and there are persistent shortages of workers in many physical jobs, including manufacturing, logistics, and care work. Humanoids and software agents are arriving in that context, not in a world of excess labor.
Physical humanoids and digital “soft bots” or AI agents will interact and reinforce each other. Many of the most advanced developments in agents are happening in the US, and those agents will help orchestrate physical work as well as white collar tasks. The challenge is not to stop this, but to manage it: helping small and medium businesses adapt, updating skills and education, and making sure benefits from higher productivity are broadly shared.
Taken together, the conversation paints a picture of humanoids as the next general purpose technology after cars, computers, and smartphones. They start in specialized industrial roles, then spread into everyday life. They reshape cities, work, and care. And they shift the balance of economic power toward the regions that build, deploy, and own them at scale.
Whether that future feels like empowerment or disruption depends a lot on what we do now.
Transcript of our conversation, edited lightly for readability
John Koetsier:
Is China going to eat America’s lunch in humanoid robotics? Hello and welcome to TechFirst. My name is John Koetsier. We all know humanoids aren’t exactly science fiction anymore. They’re picking boxes, they’re palletizing, they’re feeding machines, they’re playing soccer, they’re fighting in boxing matches, they’re running marathons.
But let’s be honest, they’re still largely prototypes. There’s a few with some paying jobs, sure, but it’s all fairly experimental at the moment. We have to admit, however, the promise is out of this world. The promise is huge, and the tech is getting better really, really quickly.
We see that China is pushing hard on scale: bigger pilots, faster iteration, more integrated supply chains. If they succeed and other nations don’t, that likely locks in some of China’s existing manufacturing advantage. How can the US and other Western nations compete?
To chat, we have Dr. Robert Ambrose. He led one of the first US humanoid teams at NASA. He’s now chairman of Robotics and Artificial Intelligence for Alliant. Dr. Ambrose, welcome.
Robert Ambrose:
Hi, it’s nice to be here today.
John Koetsier:
Super pumped to have you, super pumped to get into this. We’re going to talk about NASA and humanoids, what that was like. We’re going to talk about the future of humanoids, what they can do right now, how we’ll know when they’ve actually arrived, and then we’ll get into all the geopolitics and what it means if one country wins and another country fails there.
Let’s start with NASA. Tell me about building humanoids at NASA. Pretty cool job.
Robert Ambrose:
Well, it was, and I got to lead three generations of teams building humanoids there. And by the generations, I mean generations of the robots.
John Koetsier:
You weren’t there for 60 years.
Robert Ambrose:
So, you know, robots can replicate faster, I guess. But they all seemed to have about a nine-month gestation period to build a humanoid.
The first one we built was the Robonaut 1 system, and I was the only one who called it Robonaut 1. Everybody else just called it Robonaut. And actually the first one that we built, I called it R1A1, and people really had a hard time with that. But I was looking ahead because I knew that the way to win in this type of a technology was through iterative design. And if you try and do everything perfectly the first time, you’re going to miss. All the technology was moving so fast that I knew that the key to this was iterative, generational evolution of the design.
So the first one we built, Robonaut R1A1, is now in the Smithsonian. And it was an amazingly effective machine. I remember the first time we just had a single arm and a hand and a neck and a head and a VR station. We forever after this moment called it “stepping into the robot.” You would put on the glove that would command the hand and the arm. You’d put on the VR helmet. And again, this was in the ’90s, so we’re way ahead in VR. And you would turn it on, move your hand, move your fingers around, and you’d look down at it, and it felt like you had stepped into a machine.
And it was so evocative that you would look over and you could see yourself in the lab sitting in a chair. People would not recognize themselves. And then if you were handling an object, if you dropped it, people would move their foot so they didn’t want it to hit their toe. But you know, the robot didn’t have a toe, and they’re way over on the other side of the lab. So it was just this: you felt like you were in the machine.
And so I knew. I reached out and I grabbed something and I picked it up and it was just so… I didn’t even need training. And I just knew, you know, we’re on to something here. But what was motivating that design was a very unusual set of requirements, which was to be able to handle objects that were built for astronauts.
And the goal we set was hugely ambitious. To take it down a little bit, we wanted to have the robot have the same dexterous handling capability as a gloved hand. And a spacesuit glove is pretty debilitating. So that was… it wasn’t quite like a surgeon’s hands, but it was still quite difficult. We had a bunch of spacewalking tools, power drills, and different gadgets that were built for the human hand.
And the idea was, if people had to leave and go away for a while and there were all these tools there, how about if we just have the robot use the same equipment? And “tools” is overstated. There are latches and levers and things that were built for the human hand. And I could have asked, “Hey, just redesign everything and I can use a simpler robot.” But they would have marched me out to the front gate and said, “Go away, Rob.” We can’t start over on the design of the Space Shuttle or start over on the design of the Space Station.
So if you’ve got an environment that’s built for people and it’s very expensive, and trying to redesign it so that the robotics guys can use a simpler robot, it’s just never going to happen. A humanoid is a reasonable solution. Now, with that design, we called the Robonauts “half humanoid” because they were really only upper bodies.
And in the Robonaut 1 series, we put them on a pedestal where it was immobile. We had a waist joint. We put one on two wheels, we put one on four wheels, we put one with a single leg where it could climb and reposition itself. So we were studying different lower bodies. And that’s really where we started at NASA: thinking about the problem of working in an environment that was built for people where the hourly rate is like a million dollars an hour for a person in a space suit. Very high-end, very expensive. There’s no hope of rebuilding it to use a simpler robot. Now go: how do you go do it?
And we learned a lot in that. And I’m really proud of that team. The Robonaut 1 series: we built R1A1, and then we built R1B1, and that was it. That’s the most we did in the Robonaut 1 series. And that naming convention, yes, that is the George Lucas / Isaac Asimov four-digit naming convention. It was Robonaut series, Gen 1, batch A, unit 1, like C-3PO or R2-D2.
When we got to the Robonaut 2s, we never got to the D batch, but we got to the C batch. So with Robonaut 2, we built R2C1 through R2C6. And the Robonaut 2s were a whole other story working with GM. And it looks more like what I’m seeing today with a lot of these humanoid companies going after automotive applications. We were just way ahead of our time. That was in the mid-2000s. We were working with General Motors. They were a great partner.
The main focus there was safety around people. So the idea that you’re replacing people, that was never an issue at NASA. We just wanted it to be safe around people. And we developed some amazing intellectual property for making robots truly kind of intrinsically safe around people. Now a lot of people are selling these things called cobots, and they’re friendly. Well, we took it to a very high level of safety. And I really think that’s important. And again, what a great thing for NASA to do: figure out how to make humanoids safe.
John Koetsier:
Absolutely. And what a great thing for a career, actually, to have a product you worked on in the Smithsonian. I mean, that’s pretty cool. And there’s not that many people who can talk about that and say that. It’s really interesting. It’s also super interesting that you mostly worked on an upper body. Obviously the deployment space for this was going to be in space: in a shuttle, in the ISS perhaps, other places like that.
And you could have thought of locomotion capabilities with compressed air or something like that. You would have had to grip when you arrive where you’re going to work because you need to balance against or stiffen against forces as you turn knobs or whatever, but you could have moved without legs potentially.
Robert Ambrose:
Yes, well, you’re exactly right. When we started the Robonaut 2 project, we were working with GM, but then an opportunity presented itself to fly one to the space station. And we had built R2A1, and then we built R2B1, and that was going to be GM’s robot. And I say “was going to be” because the government, you know how the government is, we took it back.
And so we had to do something with that, and that’s why we built the C-batch — to give GM a payback for having bogarted their robot. So what we did was we sent R2B1 to the space station. And first we just sent it up as just the upper body, and it’s also in the Smithsonian, by the way. So there’s a Robonaut 2 in the Smithsonian right next to Robonaut 1 that’s in the Smithsonian. Think about that.
We learned a lot. By then, we had gone to a very high level of autonomy. For a General Motors application, having a person tele-operating a robot, that’s like the stupid idea club in a factory, right? So the robot needed to be able to do simple things on its own.
And so by the late 2000s, early 2010s, we were starting to try all kinds of amazing things in autonomy. But these are things that are more at the eye-hand coordination level, or like reflexes. And they were good for a teleoperator also, but it was more like the robot would see an object and be able to handle it, reach out and grasp it. Just basic things like push a button, turn a lever, turn a latch. We had all these — again, it was all these human interfaces that were built for people and they’re already on the space station. And if I could go back and redesign them, I could have used just a simple jog gripper or something.
But, you know, the terrible politically incorrect statement is it’s a man’s world, and it really leaves the robots out. So we just accepted, OK, we’re going to have to take the human as the design standard. And it’s hugely more challenging, but I think we’re now there. And so if you think about all the situations that are built for people, we just take it for granted.
And so at that point we started getting into other forms of mobility, locomotion. And for Robonaut 2 we built two legs. And it was more like rock climbing in zero gravity. And the legs are kind of creepy. If you look at any of the videos of the Robonaut 2 legs, they’re kind of creepy. They’re more like arms or like monkey feet or something. They’re different. They’re not like human legs.
But we didn’t evolve to climb around in zero gravity. If you look at how the astronauts do it, they’re floating. Sometimes they kick off with their feet. Humans don’t have monkey feet. So, you know, but a robot could. And so with the Robonaut 2, we designed prehensile toes that were very effective actually. So we’re not stuck with human… you know, don’t take that too far, right? Think about the job to do.
And so, flashing ahead, the third generation of robot we built…
Well, I’m happy to stop. I’m talking too much. Do you have any questions?
John Koetsier:
Well, I just want to comment right there. I just want to comment right there because you, in a nutshell, you’ve laid out exactly the argument to go humanoid or to go purpose-built for a specific reason, right?
The argument to go humanoid: it’s a human-built world and it’s all built for things that we can grip and grab and move and turn and open and all those things. And it’s built at roughly human height as well. So if you have a humanoid that walks or moves in other methods, that’s about the right height to go.
And then a lot of people are saying, “Hey, you know what, for this purpose in my factory, the humanoid doesn’t make sense,” or, “I want it to climb all these racks in my warehouse, I need a different form factor.” You’ve outlined it exactly right there. And so that’s probably why we’ll continue to develop multiple directions — humanoid for sure, but also many others also.
Robert Ambrose:
Yeah, and I think separating your thinking between the upper body, which is more like the work system, and the lower body, which is more like locomotion — which is really how to get the upper body into place to do the job — is important. Be open-minded. Just because the wheel is not something you see in nature, the wheel is actually a pretty good thing.
So we tried lots of humanoid upper bodies on wheeled lower bodies. We called them centaur configurations, which is the four wheels. And we built, for Robonaut 2, the Centaur 2 that we built. It had a bulldozer blade on one end and a humanoid on the other. And that’s a pretty good combo. You know, you use the hands for the little stuff, and if you’re not getting it done, turn around and now you’ve got a bulldozer blade.
John Koetsier:
There’s going to be some nightmare fuel in what we build for robots and what we have built for robots. I have no doubt about it. It’s craziness.
Robert Ambrose:
But your point on the purpose-built versus humanoid is such an important one, and I call it a design tension. You really have to think it through.
What we’ve got right now are a lot of single-purpose robots, and I’m a big fan of single-purpose robots. Like a Roomba that does vacuuming — a humanoid pushing a vacuum cleaner is a complicated way compared to a Roomba, right?
But here’s what I have observed in this, and it’s kind of subtle, and I’m going to make fun of it here because I’m a little sarcastic. People have come up to me — I show them a presentation of the last 40 robots I’ve been responsible for, and maybe three of them were humanoid out of 40 — and people will come up to me and say, “Why do you only build humanoids?”
And I say, “Well, what do you mean? I just showed you 40 robots and only three were humanoids. So what about that rover?” And they say, “No, no, that’s a rover.” “What about that ball that floated around and took pictures?” “No, no, that’s more like a drone.” “Well, what about the drones?” “Well, the drones aren’t really robots.” “And what about that snake tendril thing?” “No, that’s more like a borescope.” Like — dude, your definition of a robot is essentially only a humanoid. All the other things you think of as machines.
And I think it’s really that you’ve got the problem, but you don’t accept all these other things as robots. I think the modern Coke machine is a robot. It’s transparent. You can see all the bottles. It grabs the one. Ten years ago, soda machines were thieves, and now they’re very reliable and transparent. You can see what’s going on. I think that’s a robot.
What I think is going to happen is single-purpose robots will end up being thought of as machines.
John Koetsier:
Let’s pause right there. Let’s pause right there. What is your definition of a robot? What do you need to have?
Robert Ambrose:
Well, I have a very big tent. I have a very big tent definition of a robot. So I’m not the guy to ask if you want a strict definition, because I think elevators are robots. And the modern elevators work as robots.
John Koetsier:
So movement — is movement a requirement?
Robert Ambrose:
Yeah, definitely. Definitely, for a physical robot, it’s definitely a requirement. You know, the modern concept of a “bot” is a software bot, you know, that’s different. But a physical robot — there’s definitely motion and there’s definitely computer control with at least one kind of sensor and one kind of actuator. And most people would say, “It’s got to have more than four,” or whatever. I’ve got a very big definition of a robot. I mean, I definitely think a Roomba is a robot. I do.
John Koetsier:
Mm-hmm, sure. Some people think a dishwasher is a robot.
Robert Ambrose:
And I’m totally good with that. These single-purpose robots, I think we will end up thinking of them as machines. But the more general-purpose will always be a robot. And as we break out single-purpose machines, we’ll just think of them as machines. But they can run the same software architecture, and they all can talk to each other.
I don’t want to be a snob about how certain machines don’t elevate to be called a robot. I have a very inclusive definition of what I think a robot is.
But there’s one of my favorite cartoons — a Jetsons cartoon with the humanoid robot Rosie. Rosie was a maid and had a funny personality. And there’s this cartoon where Rosie is motoring by a telephone answering machine. And you could see, the way they drew it in the cartoon, you could see the answering machine was ringing. And someone says, “Hey Rosie, could you get the phone?” And she says, “No, let the machine get it.”
And I just thought, you know, that’s it. That’s crazy, right?
John Koetsier:
That’s crazy. That’s pretty cool.
Robert Ambrose:
And, you know, a general-purpose robot might interact with single-purpose machines.
John Koetsier:
I think it definitely will. Absolutely. Certainly we see that in terms of the vision of various robotics companies out there. So if you look at a Tesla, you see the vision of a humanoid that’s going to drive a vehicle and deliver a thing, right? And if you see others — Figure, for instance — you see a vision of a robot operating a coffee machine or other things like that. So varying complexity there.
OK, so we’ve talked about the past a little bit and where you’ve been, what you’ve done, and what a robot is. Let’s talk about the future. We’re going to get to today. We’re going to get to what they can do right now, what they’re capable of. And we’re going to get to all the geopolitics as well: what it means if the West develops this, what it means if the East develops this, what advantage that gives an economy or a country.
But let’s look at the future. Right now it’s pretty impressive, but it’s also pretty lame. You see the walking speed of a humanoid is nowhere near the top walking speed of a human. But you see the development, the pace of development, is quite insane, quite impressive. You see some companies — Apptronik, from the time that they started building humanoids to when they launched their first one — unbelievably quick. There’s off-the-shelf components you can get. Now you’re probably going to customize as you go deeper, but there’s so much you can do.
And the intelligence that people are bringing in, whether that’s LLMs, whether that’s other world models, other things like that, absolutely impressive as well. How do you see the future of humanoid robots in our economy, in our society, in relationship to people?
Robert Ambrose:
Well, I think the future is bright. And I’ll just tell you that we’ve got a lot more robots in our future than in our past. I mean, it’s just obvious, right? And so the question is… it’s just kind of inevitability logic that it’s going to happen. And in particular with the humanoid, there are a bunch of reasons why it’s going to happen.
And I just tell people it’s inevitable — that even if it didn’t make total sense, there’s something about the humanoid shape where people will still do it. But then in applications where it actually makes business sense, it’s inevitable. It’s just going to happen. And it’s happening fast.
So what we see in robotics is that it advances in lurches. It’s not a smooth, even anything. For example, in the ’80s, the big thing was the industrial manipulator for welding and painting. It lurched ahead. And then in the ’90s, we saw Mars robots autonomously driving, and we were seeing it in labs. And by the early 2000s, we were having off-road autonomous vehicle challenges that were doing well with DARPA. And driverless cars have come to society sooner than we predicted.
So that lurch was very uneven. It was just in driverless vehicles, and it moved to the left in the schedule. You know, when people think of a calendar, everything slides to the right — often predictions are way early. But that one moved to the left.
So why is that? Well, government investment in a Mars rover — you couldn’t tele-operate it. The round trip was 20 minutes on average to Mars. You can’t bump and wait. And then the DARPA challenge of off-road driving, that really pushed the technology. And most people credit that government investment with actually having the driverless car come earlier by about a decade.
And then we saw drones lurch. Drones went from remote control helicopters with a two-cycle engine and an amazing amount of skill required to fly them — where the hover algorithm went through the human’s eyes, where you would see the thing and then adjust — and you’d see a wind gust, and then you’d adjust. Now all of that’s running onboard a drone, and anybody can fly a drone now, because you’re just moving the character around and it moves wherever it wants to go.
So that was a lurch ahead in drones. So what are the next lurches we’re going to see in robotics? And what we’ve seen is biped mobility. Biped mobility in the last 10 years has lurched from blooper reel to running a marathon in 10 years. We went from barely able to stand to “How hard do I have to kick it before it falls over?”
So that’s where we are today. Humanoid robotics has lurched mainly with the biped locomotion. And if you stand back and squint, what are these robots doing? They’re basically, like I was saying earlier, you’re moving the hands around. And the hands are usually the part of the robot that’s going to do some work. And so the biped mobility really opens up the world to those hands — like getting in and out of a car, something you mentioned earlier.
A Roomba has no chance to ever get in and out of a car. Ever. The world we have built, assuming the athleticism of people, is just outside the reach of most robots. If you picture a little tank robot, there’s no chance that it will ever get in and out of a car gracefully. I mean, it’s just not going to happen. And so you’d have to redesign a car. And I’ve done that — I’ve worked on projects where vehicles deploy robots with a specialized thing, and you can do that. But just to go into a 1957 Chevy with a robot — you can’t do that.
And if you look at the last 50 years of cars, they’re all basically shaped for people. And there’s so much in the world that we just take for granted.
One of the wonderful challenges I got to work on was trying to help the bomb tech community. The requirement they had was just so obvious, so simple, and it has defeated almost every robot. They said the robot has got to fit between two parked cars.
John Koetsier:
[laughs]
Robert Ambrose:
How hard could it be, right? Well, in Texas, cars are… you know, a lot of pickup trucks, and they’re pretty tight. And if you look at the footprint of a robot that can fit between parked cars, the people — we squeeze in. And we don’t scratch the cars up either. Unless you’ve got some exotic belt buckle or something, you’re not going to really hurt the cars. And that’s pretty impressive.
And that’s just the world. We just take it for granted.
I see… now let’s talk a little bit about the geopolitics. OK?
John Koetsier:
Yeah, let’s get into that. Just on the parking for a moment, have you seen those parking robots that are essentially skateboards? They’re maybe six inches tall and they go under a car, and then two little arms come out for every single wheel and come close to it, and they just raise up a little bit and they move the car around. It’s unbelievable. Really, really cool.
Robert Ambrose:
Well, yeah, I think it’s really interesting, but I think there’s a simpler solution. And it turns out cars are mobile.
And so one of the really fun things we did when we were working with General Motors: they had some really deep thinking about the future of the car and how so much of our society, not only is it built around people, it’s also built around cars. And if you could redesign, if you could reinvent the car — and who could do that more easily than GM, right? That’s a big ask.
But imagine when you want to go to an office building. Imagine if we rebuilt office buildings where cars drive into the lobby. Now, they would have to be electric, because you wouldn’t want a diesel pickup truck coming into a lobby, right? But imagine they were well-behaved, quiet and clean. Imagine if you re-architected lobbies to be like indoor ingress/egress areas that could be very attractive and efficient.
And then you’d get out of your car, and then the car would disappear into a hole in the wall. And there’s some parking over there that’s robot-only. And they would pack themselves as tight as you could possibly pack them, and then they would charge themselves and maybe wash and wax. Who knows what goes on in a robot garage, right? And then when you call for your car, it comes out into this auto lobby, and you hop in and…
John Koetsier:
Clean, charged, ready to go. Sweet. I love it. I love it.
Robert Ambrose:
No walking out in parking lots, and the land use is just so much better. Parking lots are just a terrible use of land and time.
John Koetsier:
I know. It’s a disaster, exactly. I’ll tell you what: I’ll let you do that future. I’ll send my telepresence robot, humanoid, so I’ll be in the meeting too, just with my humanoid robot, and I’ll be watching through it.
OK, we’ll get into the geopolitics. You mentioned that. We wanted to get into that. It’s a big deal, right?
So we have a scenario in the world right now where — and it’s been a big deal for the last year or so — much of the manufacturing has moved away from Western nations. That’s America, that’s also Europe, that’s also Canada, that’s many countries, in fact, and moved a lot towards the East. That’s a lot of China, that’s also some other nations there.
And that was done initially because labor was cheaper, and there’s some of that as well. But now you have masses of industries that work well together, that need to be together, that need to be co-located to build a product, because you need these components, you need these components, you need these available, and it all just fits and works together. And you have that massive industrial infrastructure there.
So if we bring humanoid robots into the world, and let’s say China gets them — China’s working on them hard. They’ve got massive projects and massive investment there. The US is working on it hard too, more on a private scale than on a national scale. And so there’s 10, 15 at least companies that are significant and interesting in the US that are doing cool stuff with humanoid robotics.
But let’s say China gets there first, and let’s say no other nation gets there. All of a sudden you’ve got a massively solidified industrial backbone with almost free labor, if you look at it over the long haul as humanoids get cheaper and cheaper to build and they start building themselves, and you basically have von Neumann machines and there you go, right?
That has immense implications for geopolitics, for power, for industrial power, for wealth of nations, all those things. And we’ve seen a lot of consternation about that scenario over the past decade, especially the past year with tariffs and everything going on back and forth.
Talk about this complex economic reality as humanoids start to pour in.
Robert Ambrose:
Yeah, so the situation is actually a lot worse than you think. I’ll just start with that.
A humanoid, a general-purpose robot that can do not just one task — maybe 10 tasks, maybe hundreds of tasks — is such an incredibly useful commodity in manufacturing. OK, but it’s also useful in people’s lives.
And so the last time we saw something like that… well, that’s not true, there have been many examples. Personal computers are examples of that. Initially, computers were things that only businesses had, and then eventually there was a consumer market for them, and all the initial development was done for businesses to use computers. But when the commercial market opened up, it just kind of exploded, right?
Another example of that, much further ago in time, was the automobile. There’s a famous pair of pictures from Times Square over not even a decade. I think it was 1903 to 1910, taken from the same spot. And in the 1903 picture there was one automobile and everything else was horses — and all the things that come out of horses. And then in the second picture, you had to look really close, but there was one horse and everything else was cars.
And that was in less than a decade. And we’re talking Times Square, right? So that was where the world was turning, and it turned very fast. And if you think about all of the, like you were saying, all the infrastructure that goes into automotive manufacturing, all the second- and third-derivative kind of companies, and then all the people using those automobiles, right?
And it kind of started with trucks being used for business and a bunch of rich man’s sports — race cars, V16 race cars in 1890 or something like that, right? And so that was very expensive, high-end automobiles. And then they started making trucks and delivery trucks. And delivery trucks were really effective. They were the business version of the automobile.
And World War I — the decision to not use horse-drawn carriages in World War I was a huge strategic decision that paid off. That was a big move to go all-in on the truck versus the horse and wagon. And just think of the logistics of deploying trucks into a battlefield versus horses and wagons. And there were a lot of horses and wagons in World War I.
You know, America… even in World War II, I think it was Poland, maybe it was Czechoslovakia — I read the story of a horse cavalry that took on German tanks. It did not go well. It did not go well. So that rapid change that happened — and what happened?
Well, it got into the consumer when the following two things crossed: the vehicles were becoming more reliable and cheaper.
And a car in the 1900s had like 25 miles before you had to change a tire, right? And that was considered amazingly good. And people would travel long distances like 100 miles, with multiple tires on their rear bumper. And that was considered a breakthrough — that you could go 100 miles and only change the tire two or three times. That level of reliability, coupled with a low price, meant you could go 100 miles really quickly. And you could get a job five miles away from your house and get back and forth to the work site reliably. And you had to be reliable or else you’d lose your job. You had to count on it, and it had to be cheap enough.
Henry Ford famously said he wanted his workers to be able to afford them. And not just afford them as a hobby or a rich man’s hobby — afford them because it actually made their lives better, and it was worth the $500 they would spend buying a Model T, and they got more than $500 in value to their family having the family car.
So the price coming down and the functionality and the reliability going up.
John Koetsier:
So that’s where you see humanoids going?
Robert Ambrose:
Yeah. So as the price comes down, and especially with the onset of AI, if the ability to reprogram them to a spectrum of tasks — not just a single purpose — grows, every new task that they can do is more value. It’s more valuable that they can do not just one thing, but now they can do 10 things. And if you download the new patch, they can do 20 things.
So, price coming down, functionality and reliability going up. At some point, when they cross — when a humanoid can do $10,000 worth of value for you and it costs $9,000 — you buy a humanoid.
But what we’re seeing is more like the truck effect. We’re looking at, ironically, car companies wanting to have robots in the factory, right? And that’s bringing the price down, just like the truck brought the price down, and then eventually the consumer model, the car, came out of the truck.
And we’re going to see that. And when a family has a Rosie, and the Rosie is worth however much Rosie costs to the family, and it’s down at a price point that we’re not just talking about very wealthy people, we’re talking about average families… they buy a Rosie. And what does it do for them?
That’s when you’re down in the consumer market. Maybe it’s doing some cleaning, maybe it’s security, maybe it’s reporting on what the kids are doing, or checking in on Uncle Charlie, or whatever. Keeping an eye on the dog — you know how dogs are. At some point, if that robot is generating value to the family that’s worth $10,000 to an ordinary family, and it only costs $9,000 — and it has to be reliable. If you buy it and it breaks in a day, no one is going to buy Rosie. It’s got to be reliable and functionally valuable.
When that happens, and the price goes down and that value proposition crosses, we’re going to see a transition like we saw in Times Square — where suddenly, if you take a picture today, it’s all people, and then you take a future picture and there are going to be all these robots amongst the people. And you’ll say, “Holy cow, how’d that happen so fast?”
And a lot of the cars will be electric. Just picture what that Times Square photo is going to look like. So you take a 2025 photo and a 2035 photo. You look, and maybe you see one robot in 2025 — maybe. And then in 2035: “Wow, there are a lot of robots.” Something changed.
John Koetsier:
And it’s not that crazy far away potentially as well. I talked to Peter Diamandis about this probably about half a year ago. He released a major report, and we’re looking at prices right now on humanoids. They’re not as functional as they need to be, obviously, but we’re looking at prices in the low six figures, right? We’re looking at prices in the 100,000, 200,000, 300,000 level.
Robert Ambrose:
Oh dude, dude, they’re coming down below that. You know, Apptronik famously said they would like to go after a $50,000 humanoid. And three years ago…
John Koetsier:
Exactly. And we’ve seen price points come in phones as well. We’ve seen price points come in computers. We’ve seen those prices just get driven down. 50K is a big deal right now, but I see the 10K that you’re talking about. I see the 20K as well, not too far, not too many years in the future.
Robert Ambrose:
Yeah, well, there’s no reason that a humanoid should cost more than a car. If you look at the mass and the complexity and the sophistication and the number of motors even — the modern car has as many electric motors in it as a humanoid. It’s just made in a different quantity, which is a huge deal, right? And the car is on, like, generation 130. So that matters, right? Gen 130 is pretty good. By the time you get to 130 generations, you kind of figured it out, right?
John Koetsier:
Yeah, it is. You kind of figured out a lot of stuff.
Let’s come back to the geopolitics here, because if I take your analogies, and we take the analogy of the automobile, which got invented — we can say in Germany, I think — but got popularized and became big and reinvented a society in America first. And that was part of the industrial… I guess we couldn’t call it a revolution, but it was a revolution — an industrial awakening, especially in the post-war period, that elevated the United States to the global position and power and wealth that it has today.
Robert Ambrose:
Yes, you’re absolutely right. In many ways, the 20th century was the American century, and the car was right in the middle of it.
John Koetsier:
A hundred percent. Now with the robot, if that isn’t invented in the West, what happens?
Robert Ambrose:
Well, what happens is the famous photo that shows the transition is in Beijing, not in Times Square. Because that’s where the transition happens. And then maybe later Times Square catches up, but it’ll never catch up completely.
And why was Times Square the place that that photo was so shocking? Well, that was because that was in America and that’s where all the innovation was happening and that’s where it was happening. And, like you said, all the synergies of the first-tier and second-tier suppliers for automobiles were all in America. And there was synergy around those automobile plants with workers and the talent and the raw materials and the components, and we dined out on that for a century.
John Koetsier:
Mm-hmm. Mm-hmm. And now that revolution is nearing the end of how far it’s transformed us, and we’re entering this new one. We’ve entered multiple — there’s been smartphones, computing, all this stuff. There’s been many that have gone along the way.
Robert Ambrose:
Well, and we’ve had others — computers. But they’ve all been in America: cars, computers, personal communication phones. America, America, America… whoops.
John Koetsier:
I wouldn’t. So I’m going to ask you the question. But before I ask you a question like “who’s leading in humanoids,” I’ve already said it off the top.
China’s got huge investments here. They know they can plan at five-year periods, 10-year periods, 15-year periods centrally. That’s a positive and a negative from different aspects, but they’re going hard after this because they also have a population crisis with the one-child policy and not enough workers to fill the factories and not enough kids to take care of parents and grandparents and all that stuff. They have multiple reasons to go after this internally as well as externally.
But I will say that there are many incredible companies in the US that are looking at humanoids and building humanoids as well. The interesting thing is: will they be manufactured in the United States? And even if they’re assembled, are the components largely from elsewhere?
Robert Ambrose:
Now, it’s going to unzip. The idea of having to onshore the iPhone — that’s where we’ll be, right? Trying to onshore the “iBot,” and it’s too late. Maybe it works, maybe it doesn’t work. It’s hazardous, to say the least, that we’re going to bring computer manufacturing back to America, then phone manufacturing back to America. It’d be kind of nice if they didn’t leave in the first place.
And I think that’s the position we’re in, because I assure you that there were plenty of other places on Earth, when those famous photos were taken in Times Square, that had cars. But Times Square was the only place that saw that rapid change. The other places saw that change over 40 years. And it was in Times Square where it happened fast and in a huge and disruptive wave, because they were the disruptors. America was the disruptor. The rest of the world was the disrupted.
And you always want to be the disruptor, right? You don’t want to be the disrupted. Being the disrupted sucks. So you want to be the disruptor, because that’s where all the action is.
John Koetsier:
That’s where the power is, that’s where the wealth is, that’s where the agency is.
Robert Ambrose:
Yeah. So let me just comment a couple things on our position.
In 2006, I was part of an international study. It was a US study; we published a book — you can go read it. It was on the state of the art in robotics research. It was funded by the NSF and NASA and the NIH. And we had a workshop in DC, but we didn’t go visit all the American labs because the assumption was we already knew American labs. We were a bunch of American robotics experts.
So I went to 50 robotics labs outside the US in a year. And we kind of split up, so we covered many more than that. By the end of that year, I had my pulse on where we were. And I wrote the chapter on humanoids. And in every lab I would go to, I would go find the hidden humanoid. Literally, in Japan I could see the toes sticking out under a curtain.
And what people forget is that in that moment in the 2000s, we were convinced that we had lost robotics to Japan — that it was all over. Japan’s going to be number one in robotics and there’s nothing we can do about it. “We woulda, shoulda, coulda,” but Japan’s going to be number one, and maybe we should just count on being able to have them manufacture the robots and then we’ll buy them.
And about that time, I was talking to GM and others — who were making the most robots in the world at the time, industrial robot arms. They had policies, they were shifting their policies, where they weren’t — for export control reasons. You know how we have export control — they were asking, why were they shipping their latest robots to competitors? Right? And that was kind of a wake-up call.
And, you know, tools are important. And if you’ve got better tools than somebody else, like you were saying, it just goes through the whole industry.
So we wrote that, and that was 20 years ago. Everyone was convinced that we were going to lose to Japan.
John Koetsier:
How things change.
Robert Ambrose:
And so what we did — that led to some policy. I worked with the White House Office of Science and Technology Policy, and we brought together multiple agencies and created something called the National Robotics Initiative. And we had NASA, interestingly USDA — you know, food and agriculture has gone big-time robotics. So we had NASA, USDA, NIH.
Then we brought in NSF. NSF kind of led it because what we ended up doing was giving out grants, and the NSF is really set up to do reviews. Every cycle of reviews for the National Robotics Initiative, they brought in 350 experts that were doing all the reviews. And we were giving out massive amounts of money for the time.
And if you look, NSF did a good job watching the impact on American society. And I think because of that, the number of… and there was another thing too: we’d seen all these kids in K-through-12 all excited about robotics, and they were getting to college and there was nothing for them. So we were funding robotics at universities because the high school kids would show up and say, “Where’s my robotics team?” And there weren’t any at the university.
So we were trying to build a pipeline, and it worked. The number of robotics startups in America exploded. The number of kids working in robotics exploded. The number of PhDs in robotics… universities tend to incubate new companies. It just caught on. And I’m so proud of what we did with that National Robotics Initiative. And it’s the only reason we’re still number two. It’s just that number one is not Japan anymore.
Now the number one is China. And so, gee, what should we do? Well, again, government can help a little bit here, right? So that National Robotics Initiative, those investments — things like DARPA and DARPA challenges and Mars rovers driving on another world — these things matter. And, sadly, things like a drone war in Ukraine — that’s affecting the technology, right? And so, good or bad, government investments for peace or war can have a major influence on the technology. And so we need to think that through.
And if we do nothing, then Japan will probably pass us and become number two after… you know, China will be number one, and then it’ll be a race to the bottom for us. And that doesn’t sound like America to me.
John Koetsier:
No, it doesn’t sound like a lot of fun. It doesn’t sound like a good place to be. It’s…
Robert Ambrose:
But I am optimistic. So I want to, even though I’m concerned… and you mentioned the efficiency, I think, of a dictatorship — where they can make plans. There’s something that I think can beat a very efficient dictatorship. And it’s a little messy. It’s a little chaotic. It maybe doesn’t always move in a straight line. But it’s American innovation and entrepreneurship.
And it drives the Chinese crazy because they can’t guess what we’re going to invent next. So what we’ve got to do is unleash that. And so that’s what the National Robotics Initiative did. It helped that pipeline of talent. And those kids all started building robots and new robots and agriculture robots and healthcare robots and just on and on and on.
We’ve got to do that and make sure that we’ve got the place where the new ideas are being invented. But we have to follow through and not let us… and this was a big part of the book we wrote in 2006… that so many industries had this failure where America invented the idea and another country turned it into a commercial product and made money off of it.
So we have got to make sure that we follow these companies with manufacturing capacity to be able to do it — so it’s actually an American product. Because if it’s all made somewhere else, the margins are all going to go out of the country.
So I believe that we have the opportunity to do this. I believe the American entrepreneur and innovator can outperform a very efficient dictator.
John Koetsier:
Yeah. Yeah. Especially because that can be directed evolution in some sense, and that doesn’t work so well. I think the Chinese have learned and evolved beyond that. I think that they have allowed some chaos into their innovation process as well. I mean, the good news is you’re…
Robert Ambrose:
Yeah, but we invented chaos. It’s going to take them a long time to figure out how to be as messy as we can be.
So this is something — to bring this back to Alliant, let me just tell you. We’ve got a client base of small and medium companies, and we’ve been helping them in a lot of different parts of their company. And so we’ve gotten to know these industries. We’re working with agriculture, we’re working with manufacturing, and we’ve got a number of these sectors that are about to be disrupted, right?
And we’ve got data. We’ve been helping them get through a lot of other things where we understand their business. And so that’s something that Alliant is interested in doing: how to help these companies navigate, especially the small and medium businesses. The big companies — they’ll hire somebody to go figure it out for them. But with the small and medium companies, you’ve got to kind of work with them. And, as you know, small businesses are the messiest, most chaotic, and most important part of the American economy. Right? And so only helping the big companies is totally missing the opportunity.
So we’ve really got to do that. And that’s the Alliant history: we’ve gone after helping the small and medium enterprises, and they really need help in figuring this out.
So we’ve been talking about the physical bots, physical robots, but Alliant is even further ahead in soft bots — digital workers and intelligent solutions for the businesses.
John Koetsier:
Mm-hmm. And the evolution there is largely in the US, by the way. We’re talking agents now — AI agents — and that evolution is almost entirely in the United States. There’s some in Europe, there’s some elsewhere as well, but that evolution is huge. That’s white-collar work, largely. But that’s also going to drive some blue-collar work, and it’s going to drive some hardware, which could be robotic hardware as well. So that’s actually interesting stuff.
I’m going to go back to Peter Diamandis’ report. Now, it’s from January 2025 and the world has already changed significantly from there. But at that point, there were six major humanoid robotics startups in the United States, eight in China, one in the UK, one in Canada, none in the rest of the European Union, South America or Africa.
So the world is changing, and I’m hearing of humanoid robotics companies almost weekly that I haven’t heard of before, right? But we do have in the West some very significant and good companies. We’ve got Tesla, we’ve got Figure AI, Agility Robotics. Boston Dynamics is still in the game. They’re doing some interesting stuff. They’ve come back, away from their steampunk signs, if I can put it that way, and gone to electric actuators, right? So that’s really real.
Apptronik — I’m impressed with Apptronik. They’re doing some really, really, really cool stuff. I’m impressed with Sanctuary AI. They’ve got really interesting stuff for hands and touch.
Robert Ambrose:
Well, they’re up in your part of the woods, and I know them. I know them.
John Koetsier:
They are. They’re in Vancouver, right? And then of course you’ve got Unitree in China, you’ve got Agrabot, you’ve got Beijing H-RIC, you’ve got Engine AI, you’ve got Kepler, you’ve got Fourier Intelligence — you’ve got a lot of players in China. I didn’t even mention some of the biggest ones that have just popped out in the last half-year or so. It’s definitely a war for the future.
You made a good point actually as well when you mentioned the war in Ukraine. A lot of innovation and a lot of investment has gone to drones, military drones, defending against military drones, and you can see why, because that is a drone battle. That is drone battle number one on the planet, right?
Robert Ambrose:
Yeah, first drone war.
John Koetsier:
First drone war, exactly. And we know what the future is going to look like with drone swarms and everything like that. Hopefully that gets contained and we get a more peaceful world, but we know what people are like too, right?
But this is as critical, because that’s destruction. This is building. And it’s critical for national defense to have a strong economy.
So, super interesting conversation. We’ve got to bring it to an end somewhere, or somebody’s going to kick us off the line — I don’t know what’s going to happen here. I’ll give you last kicks: your thoughts on how you think the future should unfold.
Robert Ambrose:
Well, what I would hope is that the government comes up with some ways to work the pipeline of talent. And the companies you mentioned in China all have incredible backing by the People’s Liberation Army and other Chinese support. The American approach, I think, should be more on workforce and talent, and then maybe help in the way companies buy and deploy robots. There might be a tax break.
And I want to finish on this, though. If you hear people complaining that, you know, all the robots are our future overlords, that’s silly. If you hear people saying they’re going to take all our jobs, that’s kind of a cop-out. And I don’t believe it. We are in a moment of such long-duration, low unemployment that we need to just understand that.