Humanoid robots were once confined to the realms of science fiction think Terminator and Westworld, but recent tech breakthroughs are bringing this fantasy closer to reality. BMW, Amazon and Boston Dynamics are among big companies developing this tech, and Elon Musk is deploying Optimis, a general purpose humanoid robot which could be commercially available as early as twenty twenty six.
Investors are already jumping on the bandwagon. According to Bloomberg Intelligence, a basket of humanoid related stocks in Asia is up almost sixty percent year to date, making it one of the best performing themes this year. You're listening to Asia Centric from Bloomberg Intelligence.
I'm John Lee in Hong Kong, and I'm Kadidmitrieva also in Hong Kong.
How close are we to seeing humanoid robots being commercially available on a wide scale? What are the economic implications and what role Asia, particularly China playing this.
Name industry Today?
We have Sean Darby, equity strategist at Mizooho Securities to discuss and it's a second time on Asia Centric. He joins us here in Hong Kong. Welcome Sean, Thank you very much really appreciate you being here to walk us through this brave new world of humanoid robotics. Maybe we could just start by explaining what is a humanoid robot? How does it differ from kind of traditional robotics.
That's a very good question, I guess the traditional robots really evolved from the fact that they were labor cost saving productive machines, and principally they were deployed in the auto industry, So these were very easily to replace humans because the production line is one very long, continuous process, So most of the industrial robot makers essentially utilize these fixed robots in various tasks along the production line, and then the humanoid sort of evolved from that in two
different directions. The first one was that the logistics and warehouse industry took those fixed robots and started to allow them to have some maneuverability in placing items on shelves and taking items from shelves and putting it into another location. It was fairly rudimentary, but it started the whole process of allowing flexibility to wear a robot could be situated.
The other direction from where the humanoids developed was actually from the auto industry, so the auto industry had learned how well the production processes were using fixed robotic systems, and they were able to take a lot of different items from a car, like your sensors, your battery, and also your rotators. So the maneuverability of the auto when we would call that reduction gear and actuators. Essentially about three quarters of the humanoid came actually from the auto
industry in the last six months. The threshold for bringing humanoids into the mainstream occurred simply from the ability to access very cheap AI. So the auto manufacturers and the traditional robot makers first of all, never thought that they could purchase or acquire processing capability and pattern recognition from device from any sort of traditional software, and essentially AI came into that sort of mainstream and allowed them to absorb that at very very low cost. So you had
two things occurring. One is that the cost of purchasing AI dropped significantly. It's almost a free, open source architecture. And secondly, the sheer processing and pattern recognition, which is really what allows the humanoid to do various tasks different places, has been able to be quickly adapted into the physical structure of the humanoid. So It wasn't because of the cost of labor. It was simply because cost dropped for
building a humanoid. And as I said, it was still probably is the best example of AI being absorbed into a manufacturing process and being successfully deployed as well. So I think those two things came from different directions. The cost of AI was very low and the sheer processing power that manufacturers never thought they could attain and saw.
Why was there such an excitement Vishy.
Though, Yeah, because it's quite recently, like you've just said, six months.
Well, I think there had been in the background quite a number of robot companies, some of them which came in, as I say, from different directions, and all of them were able to leap on to this open source architecture and recognize just how quickly they could deploy a humanoid. As I said, most of the other available technology has been around for some time, but it was really that ability to combine AI with the dexterity of the humanoid skeleton.
So most of those that you had seen in the last year or so were essentially prototypes that couldn't really be mass produced. But now that you have just all the various elements coming together, and at a pricing point which allows for mass adoption. I think the way I would say to you is that you look at what's happened in the ev industry where car prices have dropped.
You get to those tipping points where suddenly there's a huge wave of demand that can actually mean that you can make these goods at a very very profitable level. And I think put it this way, there was no monopoly on creating humanoids, and because auto companies and various groups within the industry had already had some sort of manufacturing robots, suddenly you've seen the sort of plethora of various companies coming out with humanoids.
Asia Centric is produced by Bloomberg Intelligence, where more than five hundred experienced analysts and strategists work around the clock to bring you timely, world class research. Our coverage spans two hundred market industries, currencies, commodities, and industries, as well as over two thousand equities and credits. To learn more about Bloomberg Intelligence, visit bi go on the Bloomberg terminal. If you like what you hear, don't forget to subscribe
and share how good is the technology? Now see some videos of a humanoid robot that picks up an apple and gives it a customer. But when you're going to see humanoid robot actually do real meaningful things like maybe cook for you, clean the house.
So I think the way I would look at it at the moment is that to get a humanoid to do one specific task like, for example, picking apples from a table and putting it into a bowl, that has already reached the level of which those humanoids could be deployed. Where we are in transition is that you want a humanoid to do a variety of tasks and recognize when those tasks are appearing. It's sort of what's called end
to end AI. So you do one task like picking up the apple and putting it in the bowl, then you pick up the orange and peel it, and you go through a series of various different tasks. Now that at the moment is where I would say the limiting fact is in deploying very commercial humanoids, because quite clearly you don't want the humanoid just to move around this room and move the microphone and then just go back
into it one position. So the issue at the moment is really to adapt the AI so that it can do multitasks, multifunction and kind.
Of think for themselves in a way.
Yes, recognize that you've put some apples on the table and they go into one area that you need the oranges to be put in the fridge.
Done with the dog in a fridge, that's correct.
So the multitasking is where the commercial benefits from the humanoid come through. And if you look further down in terms of deployment, there are enormous productivity benefits from humanoid being deployed. Go back to your example of picking a fruit. In the agricultural industry, you have a lot of opportunities for picking grapes, pricking vegetables that are tasks that are
actually quite repetitive and very onerous. Similarly, in the construction industry you have a lot of repetitive tasks and then people have a lot of downtime as well. So if you look at sectors where traditionally it's been very hard to make productivity gains would include in the medical industry
as well, moving patients. The humanoids would be very very adaptable and very easily to be brought into those industries and function very very critically in terms of being able to produce the apples or move the patient or mix
the cement. So we're at a point now where they demand for those type of humanoids is immense in the system, and that this maneuverability with the human skeleton or you have dexterity pattern recognition and maneuverability in terms of the location where the humanoid can be put.
Is that because we talked about a couple use cases there, right, you know, so we talked about warehousing and sort of the back end manufacturing as well. There's sort of the consumer facing robot humanoid that could be in place. There's also the military capabilities, and then there's also I think the healthcare industry, right, so you mentioned transporting people. It could also be for you know, the elderly, like taking care of people in their home, especially with the cost
of care. Right now, which of those do you think will be the biggest areas for application?
So I would sort of put the military uses and industrial uses quite close together, because there are a lot of areas of industry and the military which are recognized as being very dangerous for humans. For example, in the mining industry, going into areas which have perhaps been blasted or where there's a need for a visual inspection. The human always would have that ability to maneuver and to enter.
You could also look at areas, for example, where you have had radioactive leaks or gas leaks, where again you need to implement emergency services very quickly. I would almost say the military uses come to some extent or evolved actually from these other areas, because it's actually a more natural progression again rescuing people from buildings that are maybe
in a dangerous position. I guess what I would say is that the limiting factors are really going to come down to very usual problems which are being presented with themselves as battery life, and again the capability or maintenance of the robots themselves. How long can they do these tasks without necessarily needing downtime. It'd be nice to have someone in the field doing for twenty four hours, seven days a week picking fruit, but the reality is those
robots themselves will need a certain amount of maintenance. So to answer your question is that I think you go through various evolutions, and that again the cost progression drops consistently as you go further and further down the various evolutions.
You talked about the cost reduction, how much is it right now to buy a humanoid robot?
Well, let's let's make two comparisons or three comparisons. Really, if you looked at an industrial robot that was being advertised maybe ten years ago, so your traditional one that helps to assemble cars, that price of a industrial robot has dropped by more than sort of sixty five seventy percent for some of them. So again, there's been such a large production of these items that the pricing points
have come down. Looking at the amount of competition for humanoids, I would imagine for very traditional roles such as just moving stuff around a room and so forth, looking at some of the quotes out there, you could see again price deflation coming through by twenty thirty percent. And the reason I say that is because the numbers of units being produced this year are around about one thousand for the I would say for the top ten Chinese makers.
But you know, looking out to twenty twenty six, as you start to involve the production processes, I mean we could be seeing price drops that are much much deeper than that. Again, I come back to the point that in terms of the costing twelve to eighty months ago, they never thought they could get that AI is as cheap and as abundant when they thought about what the
price point was. One other thing on pricing is that the advantage you gain for some of the assemblers is it's coming from an existing production line for the auto companies. So we've spent some time looking at where the price pressures are going to be less in terms of building a humanoid, and that's in the road tree and the actuators. So when I extend my hand, it doesn't go right through that window. There's an ability for me to control
that movement. Now, that segment of a robot was a very particular niche of engineering, and again just four or five players in that area, so they can keep the pricing quite well controlled. So I think to answer your question is that the prices look relatively static at the moment. Units like Optimus are being sort of highlighted at around twenty thousand dollars. But I would say to you, look
at what's happened to the evs B BYD. They're almost halving now, and I would say that that's probably where we go in two to three years time with some of the humanoids, that they will appear to have been halved from the price levels being offered at the moment, so.
Sort of somewhere between the cost of like a fridge and a car right now.
Yeah, I would say almost getting to the price of an electric vehicle. I think your opening remarks were very precient, which is nobody could imagine that, first of all, you could produce humanoid but at the cost levels that could be adopted so quickly. And again I point out that ironically, the auto industry was in the lead of this, but other players have come in because they've had various aspects
that they themselves are sort of masters of. Yeah, I think the auto industry did the humanoids a great favor by being able to produce a lot of the items and have a lot of the products like the batteries, the sensors all in place, so it didn't take as long to bring them humanoids into the mainstream.
So Sean, which countries dominate this industry, is this going to be another example of the US and China battling it out?
So it's a good question. I think at the moment to make like for like comparisons is a little bit difficult because many of the US players in this industry are either privately owned or they themselves are part of a technology company. And I would say, Navidia's there. You mentioned Amazon, which you could, having looked at those companies past track records, deploy a lot of capital very quickly
into those areas to bring those humanoids in. Where I sense that the momentum is with is probably with China at the moment. There are two reasons for that. One is that the output of industrial robots, which is your cash cow, is growing absolutely astonishing. I think the latest numbers for output of industrial robots in China, I think for May or for April, was something like fifty percent
year on year. These are extraordinary numbers, so you can imagine those companies are extremely cash rich now and again have squeezed out other players. So the first thing is that they've got cash cows. And the second is that China itself perhaps felt that it would never be able to access the AI as cheaply as deep Seek and so forth. So again there's been a bit of a huge shift in terms of their ability now to absorb
that and to adapt it as well. So I think the way to answer the question is that the US has a big play in this from a lot of the well known technology franchises. China's probably got the momentum behind it, and also the sheer number of different companies that have already entered that space. What I would say to you is that it's going to look like the EV industry in about eighteen to twenty four months.
Because China has over one hundred EV companies, so you think there's going to be over competition in the humanoid space.
So when we did our we went through some of the listed and private companies in China which have some form of a humanoid business. We've got to about a list of about twenty to twenty five already, and those companies now are coming to the market to raise money for the sort of you know, evolution of the production process. So on that basis, I guess one thing to say about China is that this has been also part of the made in China theme Made in China twenty twenty five.
What we're seeing is the results of that investment being put in place nearly ten years ago, and those companies that themselves having worked through and grown into businesses which again was not thought about some eight ten years ago.
I'm really curious about the economic implications you know, we talked about commercial applicability, but you know, this is part of the automation trend and automation of the labor force, and especially for the lower skilled kind of lower paid work or what we call what economists call lower scaled work, which has to be lower paid. You know, if robotics are able to take away those jobs or sort of that skills gap.
Then that enters the labor force.
So there are range of predictions, but essentially we should expect the unemployment rate to tick up somewhat unless there's like retraining program or something. But how do you think of that the economic implications, the libermarket implications.
So that's a very good question. I look at it in two ways. As a service user, so you could argue with the robotaxis that you lose the driver, but the cost experience is much much lower from moving.
A to B.
I would say for the industries that I've mentioned earlier, the productivity benefits are huge. So for the corporates, they go through a dropping the cost curve that is almost a demand side shock that they can find themselves making a much larger share of profits than they would have done from the employee side. I think there's two or three things that I would say. One is that many of those jobs are either dangerous or have physical demands which are not necessarily ones that a human being could
repeat through their lifetime. So picking grapes going through into areas like minds, so it removes some job specifications which shouldn't really be used by humans at all. In that spirit.
On the other side, one area that I think there has been resistance to has been in the medical industry because the ability for the humanoids not to be able to precisely do movements almost as to the tolerance levels you would see in building a car, means that surgeons and doctors would find themselves not able to even anywhere close to being able to replicate that themselves. Now as
a service user. That means that medical costs drop very sharply, which is a huge overhang for the global economy in the developed world, but certainly for wages of doctors and surgeons, it means a very very different transition. So again, I think the way I would say to you is that most industries have gone through similar evolutions. If you look at a Boeing seven four seven, it used to have about six or so people in the cockpit. We're down to about two or three. Now planes can almost take
off for themselves, apart from landing and taking off. So I think what's going to happen with the humanoids is that it's going to mean that a lot of those low wage, low skilled jobs are taken out from the economy. I mean, when I was young, I packed shelves in a store and it was a way of me garnering some wages. When my master I was at university, those jobs will essentially go. So maybe I'll have to work in a student union bar or something like that, or
you know, do a variety of other tasks. I think that's just going to be the nature of the evolution, really, but I would see it as being very very deflationary. All other things being equal, those areas where there's been very poor productivity would really really see, you know, extraordinary drops. Again, you could have figured on a lot of people who would never go and have surgery. Costs would be prohibitive being able.
To access that as long as the company.
Is Yeah, I mean, I think the way I would look at it is that this is the reason we have very cheap cars is because of the industrial robots. The reason we will have very cheap service prices is because of the humanoids.
Sure, you know, just getting back to your point of view, potentially working at a bar, I think you're safe, robot.
I was a pretty bad bartender, so't I'm not going back into that industry. But I feel that the good thing about these areas is that it shows that there are applications for AI and that the open source architecture that's brought this into the mainstream is going to have
a lot of benefits for society. I know there'll be some military uses, but it is really going to be a huge driver in lowering costs, and the EV industry has been a good example of that, where a large number of players force costs so low that we go through an electrification process, which in turn does a lot of good for the planet. So there's a bigger, wider holistic theme about this as well.
Perhaps that's a good note to end it on on a very optimistic note.
Thank you so much for joining us today. Thank you very much.
You've been listening to Asia Centric from Bloomberg Intelligence. I'm Katudmitreva on Hong Kong, and.
I'm joenerally also in Hong Kong. You can listen to all our episodes on Apple Podcasts, Spotify, or wherever you listen and this podcast was produced and edited by Clara Chan. Thanks for listening.