Why are you watching FSD so closely?
Because the car sales data are very transparent, and it's obvious that the
car sales have hit the wall,
whether it's in California, in China, in Europe.
So in California, the data are very transparent.
Sales are down 20-24% year over year in Q2, and California being 45% of the U.S.
EV sales in China are down 7%. That's on top of very aggressive
promotions. In Europe, it's a different story.
There's a tariff coming up. Now, the Chinese make cars go into
Europe starting this month. So we'll face a 20% increase in tariffs.
So at this crosscurrent, Tesla all of a sudden shifts its story from EV sales to
FSD. Essentially, FSD is a story of, Okay, you
buy a car, I give you a driver. So that's supposed to be a
differentiation for Tesla car to.

know, get more share in the market.
But it is our view after speaking extensively to the researchers in the
field, as well as the competition's A.I. effort, that, you know, it's a very
immature product. It takes time to get a higher quality
data, probably a rework of the underlying narrow A.I.
system and the waiting for a more CPU cluster for this to work.
Full self-driving is actually been something both had and I've been playing
with a lot lucky enough to be able to do so.
But why do you think the underlying technology, the overall scale of
investment that Tesla needs to make is so much larger than the market is
factoring in? Because this is always the long term bet
that Tesla is an AI play. Yep,
because a Tesla's addition is a little bit different from everywhere else.
Everyone else in the industry relies on sensors in addition to the underlying
A.I. software.
But Tesla wants to differentiate

yourself as a unique play in autonomous
A.I., therefore, it markets as the end-to-end
vision only A.I. play.
But if you talk to people in the area, they tell you that it's a very
difficult route because it needs a lot more extensive data input, and it
requires a much more expansive A.I. system to truly get you to L5.
So now you have this a very big operation domain design, and you have
limited hardware tool. But you want to achieve L.
So that's a very, very tall order. Speaking of hardware, we were just
showing some of the things that are top of mind for retail investors, right?
They have a platform, say, dot com. They can vote or upvote their questions.
What's missing from that is robo taxi. How critical is it that you are?
Musk gives us a date, a new date for the Robotaxi unveil.
Well, Robotaxi.

It's just the one manifestation of its vision only e2e autonomous A.I.. But Robotaxi, in fact, is easier to achieve than the FSD because there's no one behind the wheel. So in terms of the legal liability, it's much crystal clear. So if Robotaxi gets into trouble, it's Tesla that's going to follow, compensate and financially legally responsible for the accident as opposed to FSD as not as it's not fully autonomous. And the Tesla doesn't claim 100% legal responsibilities. You know, it's the insurance company needs to underwrite that risk. That's why you see lots of waymo's on the street in selective cities. But there's no Tesla robotaxi because it requires a different sort of like hardware, integrated hardware, software and AI program integration. Tesla doesn't have that. We're used to perhaps being patient on Elon's promises. He's pushed back. Robotaxi He pushed back Optimus.

Just in the last day or so.
Does that matter in the longer term if we are seeing a vision of AI?
Look, this is also an energy company that does relatively well.
Well, I'm sure energy does extremely well, but energy is not something you
give it a hundred times the price to earnings ratio.
Right. So, I mean, the stock is at 250.
The company is supposed to make of $2.50 this year.
So effectively, you are giving a 100 can price to earnings ratio.
Okay. So first you story, we'll get this to
you though, where the company where Elon Musk I want to start to go.
So the math is this. So FSD will allow the penetration to go
up to 30% of all cars sold in the U.S.. So useless, about 18 million cars.
So that gives you 6 million cars for Tesla because it's give you a car, give
you a driver.

So people will buy it for $45,000, and then the company makes a 10% profit margin. And you give three times price to earnings to that kind of a run rate. Anomalisa Ernie, just like you give to Apple, that gives you a $2 trillion valuation. That's the story they want to tell to the investor base and pacified. It might be true. It might not be true. It might be someone else. Get there first, and Tesla is not. But if you give it a 20% probability, that gives you significantly less market valuation than $2 trillion.

Tesla Stages $386B Comeback as Musk Elevates AI Over EVs

Welcome to the Bloomberg Technology YouTube channel. Subscribe for original series and mini-docs on game-changing technologies, profiles on innovators, live conversations with experts, and breaking technology news & analysis from Bloomberg TV.