THE 1 MILLION ROBOT WAR! Xpeng's IRON Aims To Beat Tesla to Mass Production In 2026

Infinity Robotic
4 followers
THE 1 MILLION ROBOT WAR! Xpeng's IRON Aims To Beat Tesla to Mass Production In 2026
Xpeng is ramping up for a high-stakes battle in the EV world, aiming to mass-produce its IRON series in 2026 and challenge Tesla’s dominance. With a target of 1 million units, the company is betting on aggressive scaling, advanced tech, and competitive pricing to capture global EV market share.
#machine #Robotics #robot
Tesla's just revealed the price of their Optimus robot,
20,000 US dollars.
Is it worth it?
And will you even be able to buy one?
Well, here's what I think.
Hello, my friends.
Welcome to the channel.
I'm Sam Evans and Xpeng and Tesla are in a race
to deliver a million robots.
Both of them wanna do it.
Both of them have set the target
of selling a million robots.
So, I mean, is Tesla really gonna sell a million
of these next year for 20,000 US dollars each?
I would say that's probably not likely.
But if they did, that would be $20 billion in revenue.
That's a lot of money.
For the past few years, Optimus has been presented
as Tesla's long-term R&D ambition.
And the truth is that Tesla, they've said many times now,
they're an AI company, they wanna sell robots,
that's more important than selling cars.
And I have to agree with them.
At the 2025 shareholder meeting,
that narrative finally began feeling like it could be real.
One of the things that Tesla says not a Tesla app
is, well, absolutely unmatched at,
is it's efficient mass production.
This is true.
Tesla makes a pretty nice profit on its cars,
which is amazing considering the prices they cost.
They've never been cheaper than what they had today.
And now they'll be applying the same logic
that made the Model 3 and Model Y so effective,
high volume, low cost production driven
by vertical integration and in-house efficiencies,
all for Tesla's most important product,
their most important product they've ever made.
Much more important than the Tesla Model S,
the Model 3, the Model Y.
This is really what will make or break Tesla.
I mean, when I say make or break,
yeah, they'll probably still be around
selling plenty of cars for the next 20, 30 years,
but it won't be their core business.
You cannot value Tesla today.
I mean, their price to earnings ratio is what,
270, something wild like that.
If you were to say that that's based on the cars
they're gonna sell, well, they'll never do.
They'll never be able to justify that.
They need to sell these robots.
They need to get them into homes, into businesses,
into Tesla's factories,
building the machines that make the machine.
This is Tesla's big moment for Optimus.
Elon and the executive team set a price target
of 20,000 US dollars for Optimus
once mass production begins,
a price point designed to shatter the market
and make the robot accessible at an unprecedented scale.
I should say though,
BYD claim they have a robot as well
that they're gonna sell next year.
They say it'll cost 10,000 US dollars.
They're gonna undercut Tesla by 50%.
Is that realistic though?
Do BYD really have a compelling robot
to sell next year for 10,000 US dollars?
My guess is no.
I think it's unlikely.
BYD is not known as an AI company.
If BYD had great software
and great efficient digitalized software,
which is what you need for a robot,
why are their cars less efficient than everyone else?
It doesn't add up, yeah?
I mean, guys, if you actually compare their EVs
to every other EV on the market,
they are amongst the least efficient.
And they're good EVs, but when it comes to efficiency,
you would think if you've got the best software,
yeah, you'd be able to make your cars more efficient.
So I doubt that BYD is gonna have a compelling robot
on sale on the market for 10,000 US dollars next year.
Will it come eventually?
Probably.
I mean, BYD, when they wanna do something,
they set their mind to it, they generally make it happen.
The linchpin of Optimus' role
in achieving sustainable abundance
and the vision of eliminating poverty
is part of what Tesla say is driving them.
So the $20,000 price point is, well,
it's a pretty important part of the equation
because if it were more than that,
most people wouldn't be able to afford one.
Here's the thing, though.
The staggering production ramp
that Tesla is planning is pretty wild, yeah?
Musk has confirmed that Tesla is aiming
for an extremely fast production ramp up,
and this is what he considers the fastest production ramp
of any complex product ever.
So what's their plan?
Here's their plan.
Part one, Optimus Generation 3 reveal in early 2026.
Generation 3 production begins in 2026.
Near-term 1 million unit per year line
at the Fremont factory in California.
Interesting they've chosen California.
2027, 4 million units a year by the end of 2027,
aspirationally.
Long-term, a massive 10 million units per year line
at Giga Texas in the distant future.
Now, honestly, this would be an interesting plan,
wouldn't it, because this would give XPeng
an enormous advantage over Tesla, enormous advantage.
It would be much cheaper for XPeng to produce robots
in China than for Tesla to produce them
in the United States.
Wages in China are much lower.
Working hours are longer.
And the other thing is the Chinese government
will pretty much incentivize you pretty heavily,
not just the government, but local governments as well.
So for Tesla to manufacture these in the United States,
they will have a significant disadvantage
when it comes to cost of production.
Yes, Tesla could do a lot of stuff
to make their cost of production more efficient,
but guys, if you've been to production lines,
I've been and seen, I've seen XPeng's car production line.
It's pretty much robots.
It's pretty similar to Tesla's production line,
not that different.
Anyway, this is not a niche project,
and clearly Tesla is planning on producing
way more Optimus units than it produces vehicles
within the next five years.
That's the reason for no new cars.
What has Tesla done in terms of new models of cars
in the last couple of years?
Nothing.
You've got the Cybertruck.
You've got the CyberCab,
which is more of a robo-taxi vehicle.
Really haven't invested,
they've got a lot of cash sitting on hand,
and they're not putting it into cars
because this is gonna make them more money, no question.
The human labor market is worth 30 times more
than the automotive industry.
This aggressive timeline is only possible
because Tesla has already spent a decade
solving 80% of the problem, says not a Tesla app.
Tesla's robots with wheels, the cars,
have served as the development platform
for the robots with arms and legs, Optimus.
To a certain degree, I think this is correct.
Optimus leverages a massive shared bin
of Tesla's existing, proven, and scaled technologies.
For core hardware, that includes actuators,
power electronics, the battery, and audio systems.
For the brain and eyes,
Optimus uses the same cameras as your car
alongside the exact same AI chip.
It also takes advantage
of some of Tesla's biggest core competencies
in manufacturing, data communication,
training, and neural sims.
This all comes together
to make Optimus's training advance extremely rapid,
building upon the same types of neural nets
that full self-driving uses to learn how to drive.
Now, obviously, learning how to drive
is a little bit different
to being a robot
and doing all the things you do as a robot,
which I honestly think a robot,
all the things you can do is far more complex than a car.
Yeah, you know, in a car,
there's a lot of complex scenarios,
but human beings are capable of doing far more
than what you can do in driving a vehicle.
By solving all these items, though, at automotive scale,
Tesla's only new engineering problems
are the bespoke robotic ones.
Elon admitted on stage at the shareholder meeting
that these are far from trivial,
calling the forearms and hands
extremely challenging to an engineer.
In fact, he went as far as calling it
one of Tesla's hardest projects ever.
By leveraging everything they learned
by building autonomous robots on wheels,
which is their cars,
Tesla's reduced a decade-long R&D challenge
into a more focused, albeit still difficult,
final stage engineering problem.
Not a Tesla app said this, right?
Interesting, while the hardware path is clear,
Elon was candid during the Q&A session
about the immense cost of intelligence.
Now, my theory is this, right?
You train a robot to do one job.
Maybe you train a robot to, for example,
be a bricklayer, build brick walls, yeah?
Then the idea is after a human being trains them
and shows them how to do it,
or a software program in combination with a human being,
something like that,
when that one robot figures it out,
you could, in theory,
download that information to the rest of the robots.
So over time, the robots will simply get
better and better and better.
However, the hardware must be safe for humans
to be around these robots,
but the AI is what makes it useful.
To make Optimus truly capable
of general-purpose tasks, though,
Tesla will likely need to spend tens of billions
on training compute,
well beyond what they've already spent
on autonomous driving.
So they're gonna need a lot of cash, yeah?
The staggering AI investment is a prerequisite
for making Optimus useful for general household
and industrial tasks,
and the challenge is going to be immense.
This cost also directly connects the Optimus program
to the necessity of the new AI5 chip,
and the TerraFab concept.
The robot's brain will be just as critical
as the key components of its body.
In fact, I think the robot's brain is far more important
than the key components of its body.
I think probably eventually any company
can figure out how to make a good robot,
but making it able to do stuff,
that's an entirely different ballgame.
Ultimately, the Optimus plan is the culmination
of Tesla's entire strategy.
The company's entire success,
its stock price success,
its actual financial success,
depends on pulling this off.
It combines the mass-market, low-cost economics
of the Model 3,
the deep technological wizardry behind four-self driving,
and a vision so grand
that it requires its own bespoke solutions
to better humanity.
But I'm curious to know,
do you guys think having these robots,
in many cases, take our jobs,
or work for us, or do things for us,
is this a good thing?
Will this better humanity,
or will it make it worse?
I mean, I think in some ways it will be better,
and in some ways it will be worse.
But either way, this is the future.
Whether Tesla does it or someone else,
it's absolutely happening.
And if you think about it, right,
the amount of effort that Tesla plans
on putting into this robot, and money,
is probably greater than any other company in the world.
I can't think of any other company in the world
that's really this dedicated,
with this kind of cash flow,
the number of engineers, the data,
or everything else that,
you know, everyone,
I can't think of anyone else who has all that.
So I would say Tesla is probably
the most likely company worldwide
to make this happen.
But I've got to say,
when I was at the Xpeng AI Day
a couple of weeks ago in China,
I was shocked by Xpeng's robot,
just how life-like it was.
I said to the person sitting next to me,
I said, that is a person.
Look at that butt.
The butt, it looks like a person,
it's moving like a person.
And other people thought the same thing.
I was intrigued to see,
oh, everyone on the internet was saying the same thing.
That's not a real robot, that's a person.
But it was a robot.
So how well are Xpeng along their timeline?
No one really knows,
but it's going to be a very interesting race.
And I'm fascinated to find out who wins it.
Let me know what your thoughts are.
Thanks for watching.
