Everyone that's a Jordan welcome back to

semi-analysis weekly. We're jumping

right into it with Nico and Raik this

week to talk everything Unitree with

their upcoming IPO, humanoid robots, and

a few of the previous robotics articles

that we've done on limit levels of

autonomy, and quadruped state of the

market. Hope you enjoy.

In terms of

how many deployments are actually real

in the real world, right? It seems like

robotics is like just getting started on

the cusp of taking over entire

manufacturing plants or

>> Yeah, not even, but yeah. I even on

their industrial deployment, I think

like

even in our our writing, I I would like

to call it we gave them a generous

framing.

>> For the fact that

their their improvements have

are very material, but

with the burnout rates, the payload, and

and like the internal accuracy accuracy

of their hardware, like

we're in baby days to put it lightly.

Um I'm actually not entirely certain

even among the partners. We're trying to

get metrics of like what this would look

like.

I think they're defining industrial

deployment very broadly. Like most of

these robots are like showing people

around

places and whatnot.

But like at best, we're talking about

the frontier of deployment is like

teleop to to pick up boxes and and stuff

of this nature. Now,

I I don't think that's true for general

robotics as a whole,

but their research playbook for them has

been huge. Like so, industrial

deployment true for robotics maybe as a

whole, pace of progress on Unitree is on

a bit of a different axis. I think the

general point we're trying to make is

it's hitting there really fast into a

more capable set of a set of hardware

form factors. Um but yeah, agreed.

There's Unitree, there's there's the

research market which is like a thing

that nobody took seriously, but then

there's there's there's robotics as a

whole is to get on the eye capabilities.

Really early days, but we are seeing it.

And then there are form factors that

obviously industrially useful. Um but

those are more sophisticated, more

expensive.

>> Yeah, like

we you know, like we weren't we aren't

trying to make

the argument so much as like hey,

broadly in industrial deployments like

this works. Like this is super viable

now.

Uh you can kind of just drop it in

anywhere. It's like not not totally the

case, right? Um like you kind of have to

think about all like the throughput and

the reliability factors.

And like yeah, like you know, in our

article we have them they still kind of

burn out quite a bit. But the point

being is that like

yeah, like the smallest task like it

kind of works, right? Like and that's

all that matters because like a while

ago nobody actually took him seriously

at all.

>> Yeah. Um so why are they going public?

What are they going to public based on?

>> Yeah, and China companies go public a

lot faster, too.

Right? Like uh sometimes it's in like

the the term sheet even of like you of

this many years to go public. Sometimes

people are like liable for the capital.

I mean you guys know this as well as

other other people do, but uh

it's a I think I think it's kind of a

necessity for them.

Um

uh

And also, I mean their their revenue

ramp is is not small, right? Uh the

market that they've carved out for

themselves is is growing, sustainable,

and they they are the the dominant form

factor for those use cases.

It's easy to use.

Uh developer kit is improving.

Um

It's easy

>> Sorry, this is a revenue thing. You say

revenue is healthy, but that's just for

being a tour guide.

>> Yeah, and but also also like research

and development and hobbyist market it

it's a bigger market than I think people

internalize, right?

>> Yeah.

>> Um and like

uh and like to be fair to um

like revenue is healthy,

but in the sense like they were uh they

were charging a lot for these robots

before, right? Like like quite a bit. Um

like even, you know, when we have it in

our in our bomb where it's like "Hey, at

27,000 pre-tax price, like these are

still 67% gross margins, right?"

Which is like absurd. Uh, and it's

mainly because

there's really not that many players

that are going to kind of like, you

know, bring down the price yet. So, it's

kind of just unitary in there gouging

the living hell out of the market while

they can. And then like, presumably, you

know, these prices just keep dropping,

but

that's like a big I think a big reason

for the revenue being so large is you

could charge before like $54,000 for

these robots that would like burn out in

5 minutes. And now it's like, uh,

you know, 30K now. So, it's like huge

drops.

They're also able to produce at scale.

Uh, they're able to produce the robots

at at a unique scale. Their iteration

cycles and their improvement on the

engineering is super super super super

fast. Such that they're trying to kind

of burn the bridge behind them a little

bit.

Uh,

in terms of Okay, if I drop the cost of

this robot, you know, and the the

American bots might be trying to go and

all the way zero to 100, you know, uh,

you know, perfect dexterity in the hand,

strong payload, uh,

you know, make all the new innovations.

And and and basically their bet is

I'm going to make an extraordinarily

cheap robot

that by the way, we're also all trying

to go over and figure out the software

for how to solve the controller, how to

do like a, you know, interesting AI

models to make robots useful.

Turns out, uh, the experimentation you

need hardware that's that's cheap,

useful, like it's going to break. Uh,

you know, you have a good service time.

Like, you know, these things need to

have experimentability.

And they've increased their capabilities

on the hardware side

as they've grown as a company. So, it's

a, you know, sort of a a bottoms-up

approach to growth in in a way that is

very amenable towards like, you know,

DJI's capabilities and etc. Which like

not the the great, you know, strongest

of drone So, it doesn't Yeah, I

>> I understand the comparison to BYD and

DJI because of the burn the bridge

behind you

sort of thing that's described, which

seems like a reasonable competitive

tactic if you want to take the whole

market and you have the ability to do

so. But

the in both of those cases like

it seemed like there was a healthy

market for drones and for cars before

those entrants came in.

Maybe EVs or like autonomy was a

different part of it, but um there were

a bunch of players. Robots is like

specifically humanoids, it's not like

there's an incumbent that they're trying

to compete against. They're pretty much

defining the market and growing with it.

Um which makes me think about solar

panels

and how China has like nine of the top

10 players in solar panels. Like if

>> Yes.

>> No.

Um

do you believe that they will have

competition in China and they're just

pulling up the banner against the US

basically?

>> I think the article's core purpose in in

all seriousness is uh is is a few

things, but one of one of one of the

core things that was the one of the main

motivators here was

trying to communicate

uh the

level of importance of economies of

scale in this market. There's obviously

a ton of AI tailwinds, right? This is

this is the obvious point. Right? Yeah,

like people happen to have bought cars

for a while for BYD, right? Like uh

EVs, you know, a different situation,

but this is a an obvious an obvious

thing that

uh

it it is separate and different. Drones

drones are similar in this regard. This

is definitely an AI demand pull, right?

Uh you know, language models have become

extraordinarily powerful economic

vehicles and tools or whatever you want

to kind of refer to them as. Uh robot

models from a research perspective of

like kind of shown their their early

signs of of of life, right?

Uh and I don't really think that's

something that anyone wants to wake up

one day and you know, say, "Oh, wow, we

missed that boat."

Right? China's been a huge leader in

industrial automation in general, right?

Their robots per worker are higher than

the US.

So, like in this is, you know, caught

more classical industrial automation,

like, you know, single picking place,

like, you know, things for like, you

know, mobile electronics, etc. But, even

their their cobots are getting much,

much stronger over the years. Getting

much cheaper. They're They're not as

reliable as some of the European and and

the US ones still, but they've improved

rapidly, right?

Uh and so, this has kind of been a

really, really big mandate in in China

anyways, like extraordinary amounts of

automation. And so, they're they're well

positioned because of what they've done

in mobile like electronics period, like

consumer electronics, and then on

automotive markets

uh to kind of really see this as a

serious thing given the markets that

they've done well in prior.

Um

So, between the fact that this is within

their core advantage already,

uh the things they've grown very heavily

in in in consumer electronics,

uh and in automotive, and the fact that

the AI talent is very, very, very clear

has allowed them to be very

forward-thinking about how to build in

the hardware ecosystem early on in the

progress the progress of the AI.

Um

>> I I kind of want to add here, too, like

uh

Jordan, I think you're you're poking on

a good one with uh the DJI comment, like

um

the fact that like, you know, DJI sold

into like the drone market, but in the

in the paper we like try and really

point this out, like

there was no kind of consumer drone

market, right? Like this it wasn't a

thing, right? You You go around in like

the early 2010s, and it's like a bunch

of like dudes in their mom's basements

building the drones for like, you know,

maybe a couple thousand dollars, or you

go and you buy like the $20,000 one that

like is like military grade, basically.

And so, it's like there was no sector

for this to begin with. And then

DJI kind of comes in,

brings out like

a pretty mediocre product now, but at

the time was like kind of groundbreaking

because it was affordable, it was

functional, it had like a camera, it

could it was like stabilized enough to

be like a useful drone for like anybody

that actually purchased one and wanted

to use one.

And then all of a sudden like

and this is where you know, it's nice to

remember my numbers. Um after like that

first release of that drone, I think

they went up to like a hundred million

in revenue. Like out of like after like

two years or something. And it was like,

oh like

>> Yeah.

>> Huh?

>> Say 130, yeah.

>> Yeah, yeah. And it was like, oh so this,

you know, like this market didn't exist

before, but like you can kind of just

invent one almost. Like show people it's

useful, it's like cheap enough, and

customers will kind of just arrive.

Um was really like kind of the DJI move

there a little bit. And like Unitree

it's you know

>> you're talking about DJI. I'm going to

throw this on screen so we can take a

look at this drone.

And the numbers you're you're referring

to.

>> Yeah, there we go.

>> 4 million to 130 million.

>> Crazy, right? Um and like it's not to

say that like, you know, Unitree is like

booming and originating the like whole

humanoid market right now, right?

Because it's so small. Like it's hard to

you know, totally declare that. But like

hey, if it happens, like

you know, we we kind of point we pointed

out a little bit here.

Um this is a pretty competent company.

They're creating robots that are pretty

useful or not pretty useful, but like

becoming moderately useful at a

reasonable price. And like in the past

this has been really successful for like

just initially getting a customer base

customer base and then scaling upward

and growing and growing and getting

better. And now it's like you know,

DJI's just everywhere.

>> I think um

it's it's not a super galaxy brain to

state to go over and say that just like

drones,

uh people had a very strong interest in

them, but they they were this,

you know, novel, extraordinarily

expensive technology that was like

basically licensed to governments or

people that could afford them.

Um

there is an a surprising amount of

people out in the market,

both within like small companies and

hobbyists that

I think again, to Rex's point, you know,

on the on the drone side is is is really

a like a a cost problem, right? And

Unitree has essentially doubled down on

the fact that when they made their

quadruped cheap enough,

people have questioned very, very

seriously like what is a quadruped

market? Now you can make the industrial

argument of the fact that uh you know,

there's enormous amounts of uh

sort of like uh

uh progress tracking, security use

cases, and etc. that I do think people

deeply underrate, but that's not the

main thing quadrupeds are selling

themselves into. People want to buy a

robot dog and see what they can do with

a robot dog, right? Whether or not these

people are extraordinarily non-technical

and are are playing with quad code and

running experiments. Uh you see social

media influencers doing it, but also

just like reasonably competent engineers

throughout Europe, the US, China, who

are just trying to see like what really

is this technology?

And if it comes down at a at a to a

reasonable price where someone uh that

can like uh on a you know, white collar

wage can go over and, you know, save up

to buy it, there there's a surprising

number of people who wanted to go over

and play with it in the same way that

you would have with the drones.

Um

I think the uh deeply peculiar part

about all of this is and and actually

this is a, you know, pretty awesome

chart to bring up around this time, is,

you know, a few years ago this was a

quadruped company.

And people really didn't care about

them, right? And I think, you know, uh I

know I know I didn't, you know, really

make this point yet, but one of the

motivations of the article is is

um

this was a quadruped company. And the

the economies of scale of like finding a

market that people just really want this

product,

they've been able to bring the cost

down,

improve their hardware, make their

engineering more reliable,

improve quality across the board, make a

product that people love, which has

given them the mandate of heaven to make

the next product.

And because they have the mandate of

heaven to make the next product, they've

been able to go up the stack of

capabilities, shrink the cost then

again, open up a wider and wider market

every time. And as they grow as a

company, they've been able to make

products that are more and more capable.

And uh,

you know, I I tell this to

pretty often as we as we work through a

handful of the articles we've already

collaborated on. He's like, "Economies

of scale is is is like China's scaling

law." Right? Like Like and we see this

time and time and time again of the

oversupply strategy really really

working well within China.

Uh,

and we've seen it several times now with

enormous businesses that started in like

funky weird ways uh,

that have surprised people by the fact

that they just don't stop iterating and

they just don't stop innovating.

>> Yeah.

>> Um

>> Let me Let me try to formulate two

things quickly. One

Yeah, sorry. One is like I want to

Yeah, I'll go first. I'll go first. I'll

go. You guys both You guys both seem

very convinced that like

demand for humanoids is just obviously

going to be there when the cost drops

and the quality improves.

>> Quality of both AI capabilities and

hardware, yes.

>> Yeah. Right.

>> This is like um

um kind of the arguments on our like

levels of autonomy paper. Going to

hearken back to this one for all the

all the readers that remember that. Um

the like point of levels of autonomy was

to show like, "Hey, you you don't need

to have the highest capability in order

to be like a useful robot.

You can be doing very basic tasks.

This is why we had the quadrupeds paper

where it's like, "Well, okay, like

Unitree has good quadrupeds, great. What

does that actually mean, right? It's

just like it's just a dog walking

around. Like, who cares?" But then like

oh, you can actually find some use cases

with it, right? Like, you can have it do

the scanning at construction sites,

which is like a very expensive job to do

all the captures. You can maybe have it

do delivery in some of the like

locations which are really bound by like

size, so it's weird to get a car into,

so it's actually economically

challenging in that case. Yeah. And so

like

um and so you don't need

I I don't my humanoid doesn't need to be

perfect, right? It really doesn't. It

just needs to be able to do like a few

things that I want it to do, and it

needs to be able to like break kind of

even on a cost basis with like another

human doing it, right? And that's kind

of where we get to that whole heat map

scenario where we show like,

"Hey, it's really not like

we're not like telling you this is a

phenomenal robot, right? Like, we make

it very clear. This is not like the

cream of the crop perfect robot right

now. But it's like, "Listen, even with

all of its challenges, even if you

assume 100% teleoperation, even if you

assume mean time failure of like what

every 20 minutes, right? And it's like 5

minutes to repair, so it's like

15 minute or I don't know, whatever.

And like you compare it to a human of it

doing this task,

it's actually just a little bit better.

Like, it's just good enough to where you

can actually put it in a warehouse and

like have it do something. And we don't

have it doing the craziest task, right?

Like, we're not telling you, oh, like

I'm moving, you know, 200 things a

minute out of a box and you know, it's

like I have to think a lot and how to

sort it. It's like no, like I'm just

taking this box

and I'm putting this box right there.

Like like very basic, but like that's a

whole job, right? Like that's a whole

job.

>> Yeah.

Okay, you're making me think of an

analogy to the ChatGPT moment 2022,

maybe the cloud code moment a couple

months ago, 6 months ago, a year ago.

Like the

it seems obvious to me that demand is

there.

Um, but we're also going through the

experience where like

there's

an indication this is going to be

incredibly useful in the future, even if

it's

just really narrowly scoped right now.

You don't think there's anything

obviously limiting

the hardware from improving on a

reliability perspective and the software

or like general operation of it

improving on a quality perspective so

that it can do more and more

economically valuable tasks more

reliably over time.

>> I think there's there's a few things to

this, right?

You know, one uh

Not not to rewind, but like on the basis

of, you know, touching on like what what

is the ChatGPT moment, right? I I tend

to take this as a bit of a misnomer.

And and the reason I say this is uh when

you know, we got you know, yeah, Sunan

3.5 or 7 or how depending on how

religious you are of like when code came

online, right? I don't like to uh talk

about like when code is solved because

code has become extraordinarily

extraordinarily more uh capable from

language models. I don't think many

people that are are serious would call

call it solved.

Um

but uh it's gotten very very useful,

right? Uh

robots, in order to deploy, need to

reach a certain level of like nines of

reliability in order to deploy. So, you

are not really a complement for very

long like an economic complement, right?

You are you are trying to replace an

individual

unit of labor. I wouldn't say you

replace a person kind of thing, but

you're adding to capacity that would be

someone who who would have done that

task, right? Now, um what does that

actually look like in practice? Well,

that might be legitimate you know

tele-up words like a cheaper person from

a from another geographic region that's

like controlling the robot that's

ideally a very high level of autonomy.

That's being you know potentially like

corrected by a person. Hopefully that

person is helping correct multiple

robots at once.

Um

And these things don't have to be just

humanoids, right? We see a lot of two

arm manipulators on wheel bases.

And it to raise point these things kind

of scale up over time.

And the hardware will improve over time.

The claim we're trying to broadly make

is that there's not that there's not a

lot of work to go,

right? People have debate on whether or

not there's going to be the tendon based

arms are going to do a Wuji and sharp

arm doing which requires enormous

precision, great machining, like

very very difficult problems to solve on

the hands still today. Whether or not

you need a hand it's depending on the

use case. We have a long way to go on

that software and a long way to go on

the hardware. In the same way when code

came online, we had a long way to go. We

had a long way to do for autonomous

research. We had a long way to do even

for things that have to do with white

collar work that's not in code and like

less verifiable domains. In the same

fashion, you will have some things come

online for robotics that like take a lot

longer due to the reliability, but you

get this interesting way to look at the

problem where you say, "Hey, uh you do

get enormous demand shocks

like that are extraordinarily powerful

that give enormous amount of capital

into the ecosystem that I think people

don't deeply internalize.

>> Can you explain the demand shock you

would you would foresee happening?

>> Uh like it like again as as the price

has come down and as the capability

increase some go-to-market opportunities

will offer enormous amounts of pull into

capital into the market, you know,

talent into the market and investment

into the customers.

>> Yeah, like you have like for example

Yeah.

>> Yeah, like like for example we we we

highlight like a few of the use cases in

the autonomy paper where it's like

you're at the point now where you know,

your kind of robot is capable on like a

few axes where it's like throughput,

reliability and like the kind of failure

tolerance of the task itself. Um where

it's like okay.

Right? In the autonomy paper my one of

my favorite ones actually is like the

cooking robot, right? Where it just uses

like the two arms, you know, to like

stir some onions, right? It's like a

okay, like super basic. I can't really

like botch that, right? I mean like I've

I'm a robot. I have a timer in my head.

I can like see how hot the pan is,

right? Like I'm not going to burn

anything. Um and so now you've got okay.

Well, the robot can cook, right? Like

how many line cooks are there like

globally, right? Like this is a huge

like pool that just opened up just cuz

the robot knew how to like stir the

onions, right? And like big big demand

shock that like Nico was saying comes

from like something like that where it's

like oh it's it's doable, I guess.

>> Well, and and and again

the the fun part also you know, shout

out Kochev.

Uh

uh I I think the the fun part about this

is

what people kind of forget in Western

markets particularly, right? Uh we're

having enormous attrition in some

industries, right? So when when Rick

points out like why do we have to be

only a little bit better than person?

Like what is that really mean, right? So

let's say the robot's still not at human

speed, right? So it depends on one what

the requirement of the task is, right?

Is it a high throughput task? Is it a

low-throughput task? Is it a high

dexterity high mix task where you're

doing a bunch of different types of

things? Or you're doing something where

the space is still reasonably

constrained, right? And uh there's

different ways to break that down,

whether it's relative to

uh if the failure mode is catastrophic.

When you consider catastrophic failures

like, you know, are you hurting someone?

Are you breaking something that's

expensive, etc.

But uh so you can kind of break down

this task in a few ways to what's ready

and what's not ready.

But

we want to look at something called the

loaded cost of labor.

Which is not just how much you're paying

someone, which, you know, if it's a

minimum wage earning job, which many of

these robot applications are not for

just minimum wage, but if it's minimum

wage, these things tend to scale

linearly over uh inflation and

regulation or whatever, but it's

relatively linear over time. But the

loaded cost of labor is something that

has been somewhat non-linear in places

like the US and Europe, where people are

just quitting at higher and higher and

higher rates. Hiring gets more

expensive. You have to do more

investment into it. And so that makes

the cost of onboarding someone, skilling

someone up,

and uh finding people really, really,

really

difficult on the business.

And so if you can get a robot that works

like half the speed, but you can get it

to work two shifts,

or maybe the speed requirement which you

get like 70% as fast as a person, maybe

it doesn't really matter anymore because

your output ends up still higher.

Or

maybe you don't need rework, right?

Maybe the robot doesn't make as many

visual mistakes. Uh

and then, you know, it might be slower,

might make manipulation mistakes, but

maybe it doesn't make visual mistakes,

and that's usually what you tend to see

as the case. So what is the unlock for

an application happens to many axes

that's very domain specific.

And so

um

when we talk about these demand shocks,

they're going to come in many forms.

They're going to come through businesses

that are entirely robot oriented, that

were service-based businesses. They're

going to come from things like cooking.

They're going to come from things like I

mean like kitting, Uh uh you know,

insertion tasks are going to come online

reasonably soon depending on like again

the level of catastrophic failure. We'll

see a lot of things in logistics.

>> Uh sorry, I'm sorry.

>> Yeah, they they like data centers is

like the the really fun one that a few

folks are going after which I'm

pretty exotic about because uh you know,

the cost of electrician. How about like

non-linear prices, right? Like you you

guys have called this out a few times. A

lot of the the AI, you know, fast

takeoffs approach are like, you know,

everyone's going to turn into an

electrician. We're all going to turn

into an electrician. Which is like, you

know, that's going to be fun. I I you

know, I got that sounds like a really

awesome way to spend my day just like

unplugging and plugging in server racks

for uh for like all of all of society

and they're all going to we're all going

to have a game competing on this kind of

stuff. But uh and and and and the thing

is is like the market for that is

massive, right?

Uh

now, whether or not it applies to the

neo clouds or not is like uncertain just

because of like, you know, what their

scale is, how fast they're building out,

and etc. But with with the massive

infrastructure build out, even on

bring-up, sure, an enormous value-add.

But people underrate maintenance and how

sticky that revenue is going to be,

right? Data centers are in sometimes

remote places. Like sometimes it's hard

to find a very skilled electrician. Uh

these are serious things.

And um

you know, the robot advantage task is

not always even just you know, pricing

on uh the labor, but sometimes it's just

an enormously value a high value-add use

case where the business really, you

know, highly rates it.

Um and they'll pay a lot for it. And

there's a lot of tasks where the

business is really really really going

to pay a lot of money for this. And

that's not just on data centers, in many

other tasks it's the case in

construction and in logistics and things

where this is true, too. And we'll see

new businesses spun up on this end where

people do things in a robot-native way,

not just an AI-native way. And I'm I'm

quite excited and for things like this.

Um and that's not just a Unitree thing,

right? You'll you'll see other form

factors, other companies.

>> Yeah, we've been we've been covering a

lot of ground on this so far talking

about

not just the humanoids, but obviously

>> Exactly.

>> quadrupeds, all sorts of robotics, which

I I think it's all all related. But

maybe we can go back to

>> Yes.

>> the one of the points from earlier,

which is like um

Well, I think there there's two things.

Uh if we're comparing the humanoid

market specifically to previous

um markets that a Chinese company has

come entered into and then dominated,

whether it's drones or solar panels or

electrical

or anything else, consumer electronics.

Um

you say that economies of scale

is like a key critical component of

that. I'm curious if you can talk a

little bit about the

like Shenzhen consumer electronics

ecosystem that powered

some of the existing ones and like how

that might come into play here.

Specifically, I'm just going to show

this picture on screen that I love from

the article where you can see

>> Yeah, it's so cool.

>> just like visually see the supply chain.

>> scale of this thing, yeah.

Yeah. It's amazing. Yeah, it's

Huaqiangbei. I'm totally botching that

name, but yeah.

>> So, I'm taking it for granted a little

bit. Maybe for the people who are

audio-only, like this is a picture of a

>> yeah, I think so this is Huaqiangbei.

This is an electronics market in I

believe Shenzhen, where it's like

I think it's a seven-story building of

just like consumer electronics parts,

right? Like the like the entire supply

chain for consumer electronics is just

in this tower. It's like unbelievable.

You can go in there, you can buy your

microcontrollers, you can buy any like

any field-oriented controller you're

looking for, you can buy any camera, you

can buy, you know, any um IMU, whatever

you whatever you want. You go in there,

you you show up with your, you know,

your yuan or yuan or whatever, and you

pay

throw it down, and then you walk out,

and you have every single part you need

to build a drone now, right? Like all

within a single building. It's like

phenomenal. It's like

just one part of kind of the whole uh

like Guangzhou, Guangdong, and Shenzhen

area, like the Pearl River Delta. It's

just like one part of it. Right? It's

like everything is in this region.

Um

Yeah. Yeah, I totally not answer your

question on the

economy of scale thing, but yeah.

>> No, but let's let's tie it into like

what I thought was the coolest graphic

in the

uh Humanoids Unitree article, which is

where you guys go through the bomb of

one of these

humanoids, right? There's arms, waist,

head, torso, legs, and you've got all of

these individual components.

And conceptually, like I can zoom in on

just one of these, like the torso, let's

say, right? And there's going to be

battery system, there's going to be a

CPU board, there's going to be a video

jetson x and x. Maybe that one's a

little bit different, but like the legs

where there's gearbox motors, joint

drivers, linkage bar, bearings.

Um I mean, if you look at all of these

uh components,

maybe this

Can you talk about like the supply chain

in

the humanoid context and how this

compares to what China has done with

consumer electronics? Like, what

components are shared? What components

already exist? Is the whole supply chain

just already done? Like, what's like

>> Yeah. Yeah. So, I think like

for I mean,

this is a bit of an overstatement, but I

think like an okay portion of this was

already kind of helped out by the

original consumer electronics market.

Right? Like, most of your kind of

standard controllers are going to come

from the like original the original

supply chain there.

It's like there's injection molding for

your plastics. There's going to be

motors, right? Which are like pretty

standardized nowadays. Different like

we'll get to the humanoid aspect, but

um point being like there's a lot of the

base components are pretty common

throughout China now. What's kind of

interesting with like

the Unitree in the humanoid case in

specific is you're watching

a kind of

ecosystem form around Unitree right now.

Or not Unitree specifically, but like,

you know, the Chinese humanoid market.

Um where

people are now making a lot of these

planetary gearboxes, right? Which is

like what goes in Unitree's arm to make

it move correctly. People are making a

lot of these now in the right spec and

size, which was like

not really necessary a few years ago,

right? Like this wasn't drones don't

really use gearboxes. They're like

mostly just kind of high-speed motors.

Um

these are like cropping up It's like

every province now has, you know,

somebody that cuts your gears,

basically. And it's like, well, nobody

actually needed these gears before. Um

and then you look around and it's like,

oh, well,

I think there was like that one article

where it's like I think there's 200

humanoid companies in China now. It's

like all these kind of keep cropping up

and like the like massive amount of

these guys showing up is like what just

eventually drives all of these kind of

new suppliers who just come in and be

like, "Hey, listen,

I got a gearbox for the low. Like, I'll

sell it to you. That's fine. Like, you

know, be my customer. We'll both do this

together. We're going to get on the

humanoid wave right now." And now like

you have this whole kind of supply chain

convergence onto like

not specific, but like very certain

architectures that work really well for

humanoids that just like doesn't exist

outside of China. Not like really not at

meaningful scale, right? And so

you're What are you kind of watching

this build in real time? And Unitree is

actively benefiting from it. Um I can

kind of

really um gush over the bomb, but I'll

I'll I'll withhold for now.

Um but well, maybe

>> maybe we can extend this a little bit to

uh

Yeah, maybe we can extend this to like

how this influences

the development of the next versions of

the systems and specifically the

reliability and thermal problems that

people are seeing right now.

>> Yeah, I think one thing to harp on, too,

is it's not just uh the mobile side,

right? Um again, we we we've mentioned

not just DJI, we not just the fact that,

you know, they have a few phenomenal uh

phone companies in the country that have

grown at like extraordinary rates over

the last decade or two.

Uh but it's it's it's their automotive,

right? So, the contract manufacturers

for a lot of these robotics companies

are the same contract manufacturers or

adjacent talent to the same contract

manufacturers that allowed the

automotive market to grow to the size it

is today, and that's why it's still

growing so massively.

I think uh to to point out to, you know,

the electronics market, the reason that

exists is cuz they have an

extraordinarily diverse ecosystem

of many, many, many players that are

small. And uh I'm forgetting the term,

that's a huge shame because the the

there's there's there's this uh it's

just like going to be misquoted. Uh but

uh there's a friend of mine uh who's

talked about this, like, you know, it's

a a sea of a thousand bosses, right?

Like, so, there's some guy who's the

best in the world at making a specific

component, who's been doing it for 20

years for your for some X number

customers who you he he knows so well

how to go over and get the great yield

on his machines, how to improve his

machines, how to go over and understand

how to make the part like perfectly and

make it super quickly, in fact, you get

it the next day. And they're all

competing. He's got, you know, some

absurd number of customers on a relative

basis to the US, and if he can't go over

and make that part at the level of

quality, precision, and reliability at

something where the unit economics work

for him, where he can continue to lower

the price, cuz they're super competitive

amongst each other on price.

That they can survive.

Uh their competition allows the consumer

like the consumer of the of the

different vendors. Uh

they all benefit from from an

extraordinarily price competitive

market. And so their quality is

basically been from the pressure of the

internal community to survive on their

own.

And and this comes from the fact that

you have several markets that are all

hardware markets that are all basically

benefiting the the level of complexity

technology that was required to build to

make things like humanoids.

Uh you know, other form factors as well,

but the fact that they have a diversity

of technology to experiment across the

spectrum is is where they get these

ecosystems from. This is not a new

ecosystem for them. It's a derivative

ecosystem.

Um I think that's the thing that people

really don't get is like

when when when our production for Apple

went over there. Uh when when when when

we built up their their supply chain

from the US, they they haven't stopped.

They've made a newer and stronger and

cheaper

products and their manufacturing

processes have created this massive

second order effect of these businesses

that have serviced all these large

companies as they've grown. And they've

become extraordinarily competitive

markets where they've just become really

really

great uh

you know, you know, a wealth of domain

and tacit knowledge that allowed them to

survive. Um so that yeah, these things

matter a lot.

>> Yeah, makes sense. Okay, I got uh I got

one last question and then we can move

to to wrap.

We we covered it a little bit at the

beginning, but maybe you can do it in

more concrete terms like

this maybe the whole theme of this

podcast so far has been quite positive

and and quite like

um

you know,

like there's a lot of excitement around

humanoids in the industry specifically.

But there are

uh

big differences between like

the deployment

versus hype right now.

Um

I know we start maybe we started out

talking about that a little bit, but

like what are the

let's say there are 100 plus humanoid

companies and inter-province competition

in China.

Like can you make the bear case for

Unitree where somebody comes along and

outcompetes them for the entire unit

humanoid market or they have less

success than you're you're currently

expecting? Like what would that look

like? What are what are the challenges

they still need to overcome, you think?

>> Yeah, I think this is this is difficult,

right? I mean I I'm not going to sit

here and say I I have a I have a magic

ball crystal kind of a crystal ball to

kind of see how this plays out in

perfect form.

The there's many players that are

emerging. You know, there's a new

humanoid company in China every week it

feels like and they're building robots.

You know, it used to take you, you know,

several months to go over and spin

something out that looks remotely okay

and this like gives a lot of credit to

the supply chain aspect of things. You

know, you'll see a new company come and

be like, "Oh yeah, in 2 weeks we just

made this robot and it's walking and you

know, it does a triple axis backflip and

all these crazy crazy things." Now,

granted

the dancing isn't that big of a deal to

be honest. Like talk about a

>> Okay, what's your what's your favorite

demo you've seen so far?

>> I I I'm boring, right? Like I want I

want things that are

hard for a robot to do, which is

repeatable precision things that require

>> onions? What What are you

>> I I mean, to be honest, like I'm a big

fan of data zunders. I'm a bit like I

want to see assembly. Like I want to put

these things put together a bike. I want

to see force and torque and

janking things around. Like I want to

see real like dexterous manipulation

that requires force and torque

understanding.

>> Right. Unitree's not on like a clear

path to having that directionally

figured out just scaling up what they're

doing today.

>> Yeah, yeah. Like what's your what's your

favorite demo?

>> The

the spring gala where they're you you

can If you watch that video, that video

kicks so much ass. I'll be honest. That

video rocks.

>> [laughter]

>> Like you go and you look at it and

they're doing like the lead like the

parkour over the boxes. That video is

incredible. I'm a huge fan of that one,

but I mean

>> dancing, don't you?

>> We We found a disagreement here on the

podcast finally.

>> It's easy for me to

It's not easy for me to dance, but but

the the robots the robots are born to

dance. Like I Like the robots are born

to dance. It's low accuracy. They can

They can around. They can land on

the wrong spot. It's still going to look

cool. Uh

and that's the thing It's just that you

know the thing is it's just it's it's

better than me. So I I I can't uh

I have to let my humanity

>> performance in in the marathon?

Or is that boring to

>> That's a really big showcase of of the

burnout not happening as as long

anymore.

>> Yeah. Yeah. Exactly.

>> That's just cool actually like

removing uh my my view like that

they could be better than me, but like

uh the fact that they can run at long

distances now it is a is a scary from a

Terminator perspective, but really

impressive from how far we've gone on

our how long our motors last. So I got I

got to give it to the guys, you know.

>> Yeah.

We're too We're too practical over here,

Jordan. But I will say

um

I do like the uh the kind of social use

cases of these things. I'm not even

social use cases, but like um

in this it's like

to be clear, I don't want this segment

to be us like, you know, making Unitree

into a joke, but I do want to point out

that like these robots are pretty funny.

Um

like big fan of kind of the the robots

where they have them walking around the

street and they're just saying

outlandish all the time.

>> [laughter]

>> These ones Being fan of these ones. Like

don't get me wrong. Uh but not exactly

taboo.

>> Yeah, people love the bots. People love

the bots.

>> Come on, you know. It's like It's like

this 4' 11" dude just walking around

like saying whatever and it does like

back not backflips, but like it's like

doing dances and stuff. It rocks. Like

I'd have this guy around.

>> I actually that that reminds me I think

I think I saw one of our colleagues ask

one on a date a few weeks ago, yeah?

>> I heard about this.

I heard about this. Shout-out to

Michelle. No, no. Um

>> Yeah, Michelle. I hope that one went

well.

>> [laughter]

>> Godspeed.

Okay, so I like I didn't mean to go on.

I derailed that a little bit.

>> Yeah, do I think Unitree's going to have

some serious competition now? Yeah.

>> 100%.

>> Uh but are they are they in position to

benefit

Look, yeah, like it's more of the

strategy, right? Unitree is

an emblem of what China has been able to

do,

and

they benefit from their ecosystem. Every

buck produced right now, even in the US,

benefits their ecosystem because our

supply chain heavily relies on them,

right? So, they benefit from themselves,

and they benefit from us currently.

>> And then the customers.

>> Yeah, is kind [clears throat] of getting

Is Unitree going to get some

competition? Yes. Are they going to be a

great business? I put my money on like

if I were going to a handshake bet, I'd

say they're going to be a good business.

I wouldn't count I wouldn't count them

out of the great research market.

>> culture.

You guys said this at the beginning,

right? This is like Silicon Valley-based

startups raising venture capital to

develop robots for all sorts of

different areas who are then taking that

capital and buying Unitree humanoids to

put in a warehouse in San Francisco.

>> Yeah.

>> And that's like maybe the biggest part

of their business right now is like R&D

for future stuff.

>> I think the warehouse that they like as

an example was a test case for us to

show that some people were attempting to

go to market with them today. I'd be

hesitant to say that that's like in any

form the majority of their use case. The

majority of the use cases, you know,

universities. I mean, granted, you know,

things are changing now with new

governmental regulations regulations in

the US. Like that will curb that quite a

bit, but it's the fact that

um

Unitree's done a really strong job of

selling to researchers who need low-cost

robots that are still usable,

serviceable, and strong

uh like a good you know good to go over

and train AI models with.

Um, I am by the way I I do say this very

seriously. I I am bullish the US in

being able to compete in this market.

It's going to be very hard.

But, the US does take this increasingly

seriously to make sure that we you know

have

some form of our own supply chains is

very very very difficult. I think we're

under invested massively. We don't have

you know metal processing. We don't have

our neodymium production is too low. We

all the chemicals if we wanted to set up

processing plants are still going to

come from China.

But, we don't make a lot of PCBs here.

We we don't know how to go over and make

it we can wind our actuators, but

they're you're still going to you know

have them produced mostly in China. Like

these are really really big problems.

But,

uh the the necessity as the AI

progresses is going to just I I you know

encourage us to have massive massive

massive investment. Um, right now

China's just in a phenomenally

advantageous position for themselves.

And and this to me this article was was

was also a tending to be a wake-up call

saying like you you don't have to

believe that we are at you know you know

what Optimus's dream is what Figure's

dream is for having like the

full-fledged you know you know zero to a

hundred level humanoid to go over and

show that this market's moving. And the

point of this article is that this

market's moving. Uh the technology

is showing signs of life in a unique way

and the supply chains are very very

clearly feeding into that increasing

attractor state over time where their

advantage is not going away anytime

soon. So, to look at them like a toy

company when they were just a quadruped

company a few years ago,

I wouldn't ever count these guys out

because they've already shown enormous

progress.

I mean I'm sure by right we're going to

see a lot more from different companies

in China, but we won't we won't

US is going to put quite a bit of effort

into this as well.

>> Yeah, it's the fact that they're even

getting into this market at all is like

significant, right? Like this wasn't

even this wasn't in conversation a year

ago basically. I do want to like

I want to I want to harp on like one

thing here because this is a bit more

anecdotal and I think in the piece it's

really one part of it's not in the piece

and the other part is in the piece but

it's a bit highlighted or it's a bit

like looked over is like

the kind of scaling of the Unitree like

pro like the Unitree project as a whole,

right? Like

we mentioned, you know, they came into

quadrupeds. They go boom, okay, great

quadrupeds work. 95% cost drop.

We're shipping, you know, tens and

thousands of these things now.

Excellent, right? Like this works number

one. But we like really wanted to kind

of we couldn't draw a perfect one-to-one

correlation. So like the paper doesn't

say this explicitly, but in my heart of

hearts

um

the kind of idea was you

get your actuator for the quadruped good

enough, right? And a lot of the systems

designed for the quadruped good enough,

right? A lot of the actual mechanics,

what you're putting into the robot, all

the parts, you get this good enough, and

then you can just kind of transfer a lot

of the like actual components toward a

humanoid, which is what technically the

original H1 was. And in the in the

piece, we mentioned this, super cool,

such a good data point, like I love the

guy for telling me this. There was some

some guy close to Unitree,

but the the H1 was originally designed

as a quadruped standing on two legs,

right? You like you look at it and it

it's so crazy. Like you look at it and

it's it looks bizarre, right? It's like

the legs are already bent like the

quadrupeds are. You look at it walk and

it kind of does this thing where it

patters its legs a little bit. It's it's

designed to be a quadruped on two legs

basically. And so it's like you go from

that and then you look at like then we

can like look at the S1 where they

shipped like what was it? Like 400 in

the beginning of 2025.

And then

it suddenly jumps to like 4,000 like

over the next 9 months. And then 3

months later in January they go, "By the

way, we're actually we're at 6,500 right

now." Right?

Like visibly seeing the scaling

happening in real time. It's

unbelievable what they're doing right

now. Like just it's early stages,

obviously, but like this is super

impressive. Yeah. Okay, guys, we've got

to move to Rob here, but this has been

great.

>> Anything you think is uh left unsaid?

>> Um

I hope uh

>> early stages are here. Like

I I think I think uh it's it's no longer

sci-fi and I I think it's going to get

increasingly weird over the next like 4

or 5 years. Like uh

>> First inning. Play ball.

>> And uh

yeah, game began. The game began, all

right.

>> Would you would you would you get a G1

in your house, Nico? Or would you get an

H2 in your house?

>> I don't Yeah, again, like I'm a little

security concerned. Like I do think this

is real.

Uh I I I I do think the geopolitics is

going to get very peculiar. Uh

the US needs to prioritize this kind of

stuff. Uh

we need to figure out our own supply

chains. Unitree is a phenomenal company.

How how these things develop is is is

going to be, you know,

uh

Yeah, I'd to put it lightly.

>> Uh would you would you get one, Rick?

>> I

you know, it'd just be kind of freaky to

look at it in the hallway. That's the

only that's my only problem, right? Like

I wake up to get a glass of water, it's

like this [laughter] 6-ft dude like

>> I got that side by door.

>> Start with a roommate.

>> Yeah, [laughter] yeah, I'll get

>> If you can't handle a robot folding your

clothes

>> I want a robot folding my clothes. I'm

pretty excited for arm robots.

>> Yeah, I'm I'm I'm I'm excited for the

clothes folding robot, you know, it

doesn't need to walk and look like me,

but I could do a close folding robot.

That's fine.

>> Put a Hawaiian t-shirt on it, it'll be

fine.

>> Yeah, yeah, put some sunglasses on him,

call it a day.

>> What? One for the one for the floor in

the apartment complex? So it just goes

door door-to-door and folds your clothes

every now and then? You don't have to

>> I'm I'm pretty excited about it. A robot

as an amenity. Like I actually like I've

been waiting for this from me for years.

Uh I I I'm pretty excited a robot as an

amenity.

Like uh you know, hey, like my my my

apartment complex has a gym. Like yeah,

my mine has a robot that does all my

stuff. Like that's going to be pretty

cool.

>> [laughter]

>> Give the clothes to the robot.

>> Keep going and say like 100% to do that.

100%.

>> Yeah, makes sense. Okay, guys.

Thanks for joining the podcast today.

Appreciate it. Nice job.

>> Nice job, guys.

>> Thanks, guys.

>> Have a good one.