The ocean was on my right. Mountains to my left. My car was cruising through the curvy California freeway from Los Angeles to Palm Springs.
Then, I took my hands off the wheel…
I found myself itching to grab for it as the car headed 70 mph into the first turn.
Almost like magic, the car gracefully navigated the turn. It even slowed down to keep a safe following distance from the car in front of me.
Even though I put my hands back on the steering wheel, I didn’t need to touch the gas or the brake for nearly two hours!
This was my first time driving a Tesla Model 3.
That was seven years ago, but the memory is still vivid in my mind. I could see the future of autonomous driving clearly laid out in front of me.
Yet, when I wrote about the robotaxi’s potential to save Tesla (Nasdaq: TSLA) earlier this year, not everyone agreed with me.
One of our readers, Steve, wrote in to say:
Please look at the difference between Waymo and Tesla. I’ve owned a Tesla since 2018, and have been riding in Waymo’s for the past year. Tesla is nowhere close to providing cars as robotaxis. I know you believe it….but Teslas tech just isn’t there to provide safe rides.
Now, maybe Steve is right…
But that didn’t stop Elon Musk from launching his robotaxi service this month.
And it’s already paying dividends for Tesla.
Tesla’s Robotaxi Hits the Streets
As Musk suggested months ago, Tesla’s long-promised autonomous ride-hailing service quietly went live in Austin, Texas, last Sunday.
It started with a small fleet operating in limited conditions. The ride area was geofenced, and a Tesla employee rode in each vehicle as a safety precaution.
But that doesn’t change the fact that customers have now begun hailing rides in Teslas with no one behind the wheel.
It’s the first public rollout of Tesla’s Full Self-Driving system as a true commercial robotaxi.
And it’s a significant moment not only for Tesla, but for the widespread adoption of autonomous vehicles.
Because Tesla’s robotaxi approach is unique in this still-developing industry.
Unlike rivals Waymo and Cruise, Tesla doesn’t rely on lidar sensors, pre-mapped routes or complex external hardware.
Instead, it uses a vision-only system…
A network of cameras, neural nets and Tesla’s custom-built Dojo supercomputer, which is an AI training system designed to process video data at a massive scale.
According to Tesla, Dojo can train models using millions of video clips from the fleet without needing labeled datasets or hand-coded instructions.
In essence, it lets Tesla teach its cars how to drive by watching themselves drive.
This is what Musk calls an “end-to-end AI” approach.
Instead of breaking down the act of driving into dozens of separate tasks — like identifying objects, planning routes or reacting to events — Tesla’s AI is trained to handle everything in one continuous learning loop.
Think of it like teaching a single AI brain to drive instead of creating a classroom of separate AI specialists who each know a specific function of driving.
Naturally, some experts are skeptical of this approach. They argue that Tesla’s method is less safe and harder to validate.
And it’s true that Waymo’s robotaxis, which launched in Phoenix and San Francisco years ago, have logged millions of autonomous miles without relying solely on cameras.
Waymo trained its systems on 500,000 hours of driving.
But that doesn’t necessarily mean Tesla is playing catch-up.
You see, Waymo just published new research showing that performance in autonomous driving improves reliably when you increase model size, data and compute.
In other words, it found that scaling works…
Just like it does in large AI models like ChatGPT.
But this is exactly what Elon Musk predicted years ago. He argued that the more real-world data Tesla could collect, the smarter its system would become.
Tesla has access to far more real-world driving data than any of its competitors.
The company has logged over 100 million miles of driving data from human and AI-assisted trips.
Which means if Tesla’s end-to-end system can take advantage of this scaling law, it could eventually blow past systems that depend on pre-mapped environments or costly sensors.
It could also make Elon Musk even richer.
He’s on record that Tesla’s robotaxi fleet could one day contribute between $5 trillion and $10 trillion to the company’s market cap.
And I believe Elon might be right… eventually. But it will take time.
Still, over the next few years I expect we’ll see massive growth in this industry.
ARK Invest projects that autonomous ride-hailing could become a $951 billion global market by 2030.
That’s nearly the size of the global airline industry.
Here’s My Take
I get why reader Steve might still be skeptical about Tesla’s robotaxi rollout.
He has every reason to be.
After all, we’re still in the very early stages of the launch. It’s limited in geography and scope, and it could take a while for this service to become profitable.
But what we’re seeing now is proof of concept that Tesla’s unique AI-first approach is working well enough to be tested with real customers.
One analyst estimated that robotaxis could add $2,000 to $3,000 in annual recurring revenue per vehicle.
That kind of upside could turn Tesla from a car seller into a service provider, with steady revenue from high-margin rides.
And it’s already helping lift Tesla’s stock price.
At yesterday’s close, TSLA was up over 7% from last week’s low.
Source: Yahoo Finance
And even though Tesla faces a lot of challenges ahead to legitimize its robotaxi fleet, there is a path forward.
If Tesla can continue to improve its system by leveraging the same scaling laws that drive AI breakthroughs, the company could easily catch up to Waymo…
And maybe even leave it in the rear view mirror.
Regards,
Ian KingChief Strategist, Banyan Hill Publishing
Editor’s Note: We’d love to hear from you!
If you want to share your thoughts or suggestions about the Daily Disruptor, or if there are any specific topics you’d like us to cover, just send an email to [email protected].
Don’t worry, we won’t reveal your full name in the event we publish a response. So feel free to comment away!
