A few months ago, I asked Dan Ives when I was finally going to get the robot I’ve been waiting for since I was a little kid.
His answer was three years.
I’ve already written about why that answer completely changed the way I think about humanoids.
More recently, we talked about The Robot Stack — all the chips, sensors, motors, software and other stuff that has to go inside these machines before one of them ever makes it into my kitchen.
But this week Elon Musk said something that made me look at this whole thing differently again.
He thinks there could be at least one billion humanoid robots within 10 years.
One billion.
Now, Elon has made some pretty aggressive predictions over the years. Some happened. Some took a lot longer than he expected.
So I’m not particularly interested in arguing about whether the number will be 300 million, 700 million or exactly one billion.
There’s also something pretty funny about publishing this on Labor Day.
We’re supposed to be celebrating the American worker today, and here I am writing about the possibility of a billion robots joining the workforce.
Happy Labor Day, everybody.
Nobody said all the labor had to be human.
The interesting part is how he thinks we get there.
Eventually, the robots start building the robots.
Wait, the Robots Make More Robots?
Think about how different that is from almost anything we’ve ever manufactured before.
At first, humans build the factories. Humans install the equipment. Humans assemble the robots. It takes workers, money and time to increase production.
Eventually, you reach a point where some of the robots coming off the assembly line can go directly to work helping build the next robots.
Then those robots help build more.
This is where the math starts getting weird.
We’re used to thinking about factories as having limits.
If you want to produce twice as many cars, refrigerators or washing machines, you generally need more workers, more factories, more equipment or some combination of all three.
What happens when one of the things coming out of the factory can immediately become part of the workforce inside the factory?
Now production itself starts helping increase production.
Remember, these aren’t machines that only know how to weld the same piece of metal in the same exact spot 10,000 times.
That’s the kind of industrial robot we’ve had forever.
Humanoids are supposed to learn.
That changes the equation.
If I spend five years becoming really good at something, one person got better. Congratulations to me.
If a robot learns how to perform a task more efficiently and that improvement can be turned into software, potentially the entire fleet gets better.
You don’t send every robot back to school.
You update them. Now they all know.
That might be the most important part of this entire story.
We’ve spent the past few years watching artificial intelligence get smarter at an incredible pace.
Now we’re giving that intelligence eyes, hands and legs and sending it out into the physical world.
The AI boom was about giving computers better brains. Humanoids give those brains bodies.
Eventually, those bodies may help build more bodies.
That’s a different kind of industrial revolution.
So What Are We Supposed To Buy?
This is the part I care about.
Whenever a story this big comes along, Wall Street immediately wants us to predict the winner.
Which humanoid company wins?
Is it Tesla (TSLA)? Figure (FIGR)? Somebody in China? Some company that doesn’t even exist yet?
I have absolutely no idea.
And I think pretending we know today is probably the wrong way to approach it.
Think about the internet in the 1990s. The big idea was obviously right. The internet did change everything.
But knowing that the internet was going to change the world and knowing exactly which internet stocks to own were two completely different things.
Same with smartphones. Same with artificial intelligence.
The theme can be right while your stock is wrong.
That distinction is important.
So instead of trying to predict which robot is going to be making my coffee in 2035, I want to watch where the money is actually flowing.
Which semiconductor companies are outperforming?
Which sensor companies are breaking out?
Which automation stocks are making new highs?
Which businesses inside the Robot Stack are seeing their earnings estimates rise while investors keep buying the stocks?
That’s our advantage.
We don’t have to know what the world looks like 10 years from now.
We have price.
If humanoids really are becoming one of the biggest new sources of labor on Earth, I expect the companies benefiting from that trend to start telling us long before there are a billion robots walking around.
They’ll tell us through their stocks.
Some will break out. Some will outperform. Some will become leaders.
And plenty of others with great robot stories will go nowhere.
Those are telling us something, too.
That’s why I don’t want to fall in love with the story.
I want to follow the trend.
Because imagine for a second that Musk is even directionally right.
Forget one billion robots.
What if there are 100 million?
Every one of them needs computing power. Every one needs sensors. Every one needs motors, batteries, bearings, actuators and all sorts of other components most people have never heard of.
Then they need electricity, factories, maintenance, replacement parts, and somebody has to make all of it.
That’s where I think the opportunity is.
We already laid out The Robot Stack because I think looking underneath the robot is more useful than trying to guess whose logo eventually winds up on its chest.
Now I want to go one step further.
I want to find the parts of that stack where demand starts accelerating and price confirms that investors are figuring it out too.
Maybe it’s robot vision, actuators and possibly semiconductors.
Maybe it’s some boring little company making a part inside the hand that nobody cares about today but suddenly becomes essential when the world needs 50 million robot hands.
I don’t know yet.
That’s the whole point.
We don’t have to predict it.
We can watch it happen.
Because if robots really are going to start building robots, somebody is going to make a fortune selling them the shovels.
Our job is to figure out who.
Stay sharp,
JC Parets, CMT
Founder, TrendLabs