Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

[Ellec]

reminder, Troy will adjust his estimates Tue-Wed. Stay tuned.

Not sure how valuable estimates are if they change all the time as new information comes out, why not just wait on Tesla releasing the figures?

 

[CaliBear1]

Been thinking about "the moat" lately as FSD gets incredibly good we are marching the 9s.

What is the FSD moat? A fleet of millions able to collect video/map/weather/environment/vehicle dynamic/cabin driver data capable of being turned into ground truth for NN training that is distributed out to the entire fleet via OTA anywhere with a cycle time of less than 2 weeks.

With infinite money (100B+), how long would it take to 'catch up' to this? Well, never, why? Let's break this down and I'm leaving out the pesky parallelized bits like standing up and building out your multi-exaflop NN infra.

You'd need a fleet of vehicles with Tesla's capabilities. Does this exist today? No. Okay, so how long to build it?

You'd need to re-design your vehicles that you are producing today. Um, I'm not making cars today. Great, then just start from scratch on the important techy bits. Okay, how long does that take? Well, how much is different? The entire communications infrastructure, all the ECUs and sensors that process and send data, full bring up and stack development on vision NNs which nVidia can sell you somewhat. Okay, do I currently have engineers that can do this? No. These are folks you'll need to hire. Okay, let's hire them! Well, these folks are super expensive and you'll need to figure out how to get them to agree to work at your company. Okay, I have money, great, then let's get started! And that my friends, is the story of Project Titan at Apple.

There's another story here that is currently playing out however and that is licensing. This story is yet to be written, but I feel that is DOA and here is why.

I've got a hankering that Tesla is about to solve FSD, so I think to myself, self, how can I get that tech working on my vehicles. And let's say I can convince Tesla to provide me a giant spec book of everything I need to do. Wonderful, now I just need to start making those changes. How long will that take? In the past I've been able to make brand new vehicles in 3 to 5 years. But those cars were largely just the same techy bits with new ICE engines. Ugh, so I'm back to hiring techy bit folks who are super expensive. Gotcha, need a boat load of capital, a super focused VP and a CEO who is all in. This is what I think is going to happen with Ford as Doug Field (my former manager at Tesla) and Jim seem to walking this path in the coming year and the earliest model year to support Tesla FSD would be 2030 at the earliest. And if they don't start this year, the earliest would be 2031.

But I think this is moot, assuming Tesla builds/completes at least 4 more gigafactories by 2028 and saturates the market with robotaxi's. But yeah, quite a moat.

Also, I should mention, no other company is pursuing vision only at scale. Competitors in China are the only thing that could eclipse Tesla for manufacturing and AI, but that is just in China, the race is now getting Europe robotaxi ready and India completely built out for manufacturing and robotaxi's.

In summary, the moat makes it moot.

Wow! Thanks DiscoDucky! Fascinating perspective from a former insider that I suspect will help a lot of anxious shareholders ride out what could be a stormy 24Q1!

 

[Ellec]

Good point on vision only. Really now that FSD12 looks so good, any 'competitor' who mentions Lidar should just be laughed out of the room. Until competitors understand that vision-only is the way, they cannot even START to compete.
Honestly the biggest threat to Tesla FSD may well be regulatory. Lobbyists may insist that this tech is in all cars, and insist Tesla license it at affordable prices.

Tesla will gladly license it, however it will be direct to the customer and only available in a Tesla vehicle. Once it's proven to be x times safer than human driving people will pay through the nose to reduce the risk of injury/death to almost 0 when in a vehicle.

 

[joh01652]

It does not have to increase revenue by much initially. The very notion that this is becoming a product with appeal outside the Tesla/tech enthusiast community is a big deal.

This is my thinking, too.

Yesterday I volunteered at a delivery center giving 12.3.1 demo drives in a 2024 Model S to people picking up new vehicles. People are very impressed (even I was impressed as it was my first experience with vision-only auto park). Car worked very safely on every drive and parking session, no panic/unsafe moments in a 8 min round-trip drive in moderate traffic ranging from 25 mph-45 mph streets and parking lots with one round-about. One gentleman, who got his first EV, a CT, video recorded the drive and was very impressed. Another couple, who I don't think ever considered purchasing/subscribing to FSD, were going to go home and give it serious thought/discussion. Everyone had a "huh" moment, like, "Huh, this is something I've never experienced and didn't know existed." Moreover, only one person knew they were getting a 30-day free trial of FSD. The guy with the CT is bummed he isn't able to get it--who knows, maybe he'll buy another Tesla with FSD now if he was impressed enough. At least he'll be thinking about it, and I'd bet $100 he will buy/subscribe to it once FSD is enabled on CT.

Of the 20 or so people I chatted with buying their first Tesla, all had questions about very basic things like preconditioning the battery, premium connectivity, home charging, supercharging and much more. We really do live in a bubble here on TMC with regard to all the features and capabilities of Teslas that we know about. The vast majority of people have no clue, even those who are purchasing right now.

The new FSD v12 is crazy even for me. It really is so so much better, and a driverless future feels much more real now. I'm with Cliff Harris in my thinking that this can't be easily replicated. Maybe there is another company that is doing what Tesla is doing, but not here in the U.S. No other U.S. company has this kind of data from random people employing hardware/software across the globe in diverse and challenging real-world scenarios. Tesla is in a class all by itself. If we got news of even one legacy auto company licensing FSD tech, the stock would rip, IMHO.

 

[DarkandStormy]

reminder, Troy will adjust his estimates Tue-Wed. Stay tuned.

He always puts out his final estimate on the last day of the quarter. Not sure why you're hating on the guy, but if you're going to make a joke about someone else's work it should probably at least be based on reality.

 

[Krugerrand]

reminder, Troy will adjust his estimates Tue-Wed. Stay tuned.

Waiting with bated breath.

 

[Krugerrand]

Good point on vision only. Really now that FSD12 looks so good, any 'competitor' who mentions Lidar should just be laughed out of the room. Until competitors understand that vision-only is the way, they cannot even START to compete.
Honestly the biggest threat to Tesla FSD may well be regulatory. Lobbyists may insist that this tech is in all cars, and insist Tesla license it at affordable prices.

Cue The Knight guy.

 

[Krugerrand]

Not sure how valuable estimates are if they change all the time as new information comes out, why not just wait on Tesla releasing the figures?

Pretty sure he was being sarcastic considering Tesla will be reporting the numbers at the exact same time he mentioned the guy would be changing his estimates.

 

[Hiline]

I am not sure this is completely accurate. While it is true that tesla does have 6 million cars delivering data, the reason the jump from version 11 to version 12 was so substantial was because of AI / end to end Neural nets that tesla is using now. As we now know, a neural network is a method in AI that teaches computers to process data similar to a human brain. So now, tesla is “feeding” these human images for the AI/Neural nets to learn. And, Nvidia “bet” correctly on the GPU’s and chips to create and help this situation. So, it is almost like tesla went from version 1.0 before to 2.0 now. It is true that there are no car companies that have all the real world driving that tesla has, however, the end to end neural nets/AI learns so quickly that the playing field has been narrowed some. Tesla still has a lead, but this new AI compute GPU’s and chips that any car company can buy only gives tesla a smaller lead and other car companies can buy their way there, so maybe only a 1 to 2 year lead?. I did watch a video recently on this (seeing if I can find it) but there are videos on YouTube that discuss this, including videos from Nvidia discussing their AI tools for autonomous vehicle developers That people can view.

The notion that AI learns quickly is funny. Anyone who has built a neural network knows that it takes millions or billions of observations to develop a somewhat accurate image recognition algorithm, and yet the more data you feed the algorithm the more accurate it becomes. The reason data is the new oil is precisely this.

Architecture aside, if no competitor has even 10% of Tesla training data, it will take them years or maybe forever to catch up on neural network-based autonomous driving.

 

[2Pearls]

Been thinking about "the moat" lately as FSD gets incredibly good we are marching the 9s.

What is the FSD moat? A fleet of millions able to collect video/map/weather/environment/vehicle dynamic/cabin driver data capable of being turned into ground truth for NN training that is distributed out to the entire fleet via OTA anywhere with a cycle time of less than 2 weeks.

With infinite money (100B+), how long would it take to 'catch up' to this? Well, never, why? Let's break this down and I'm leaving out the pesky parallelized bits like standing up and building out your multi-exaflop NN infra.

You'd need a fleet of vehicles with Tesla's capabilities. Does this exist today? No. Okay, so how long to build it?

You'd need to re-design your vehicles that you are producing today. Um, I'm not making cars today. Great, then just start from scratch on the important techy bits. Okay, how long does that take? Well, how much is different? The entire communications infrastructure, all the ECUs and sensors that process and send data, full bring up and stack development on vision NNs which nVidia can sell you somewhat. Okay, do I currently have engineers that can do this? No. These are folks you'll need to hire. Okay, let's hire them! Well, these folks are super expensive and you'll need to figure out how to get them to agree to work at your company. Okay, I have money, great, then let's get started! And that my friends, is the story of Project Titan at Apple.

There's another story here that is currently playing out however and that is licensing. This story is yet to be written, but I feel that is DOA and here is why.

I've got a hankering that Tesla is about to solve FSD, so I think to myself, self, how can I get that tech working on my vehicles. And let's say I can convince Tesla to provide me a giant spec book of everything I need to do. Wonderful, now I just need to start making those changes. How long will that take? In the past I've been able to make brand new vehicles in 3 to 5 years. But those cars were largely just the same techy bits with new ICE engines. Ugh, so I'm back to hiring techy bit folks who are super expensive. Gotcha, need a boat load of capital, a super focused VP and a CEO who is all in. This is what I think is going to happen with Ford as Doug Field (my former manager at Tesla) and Jim seem to walking this path in the coming year and the earliest model year to support Tesla FSD would be 2030 at the earliest. And if they don't start this year, the earliest would be 2031.

But I think this is moot, assuming Tesla builds/completes at least 4 more gigafactories by 2028 and saturates the market with robotaxi's. But yeah, quite a moat.

Also, I should mention, no other company is pursuing vision only at scale. Competitors in China are the only thing that could eclipse Tesla for manufacturing and AI, but that is just in China, the race is now getting Europe robotaxi ready and India completely built out for manufacturing and robotaxi's.

In summary, the moat makes it moot.

Thank you. And thanks to @cliff harris , @EVWatcher , @Singuy , @CaliBear1 @Tony73 and @Hiline for all of your thoughts on this topic. All were very helpful and informational

 

[Discoducky]

Good point on vision only. Really now that FSD12 looks so good, any 'competitor' who mentions Lidar should just be laughed out of the room. Until competitors understand that vision-only is the way, they cannot even START to compete.
Honestly the biggest threat to Tesla FSD may well be regulatory. Lobbyists may insist that this tech is in all cars, and insist Tesla license it at affordable prices.

Well, if that happens, Tesla has already won not only the majority of US vehicle market, but assuredly Europe as well and that sounds like a win. And I do expect that after the first billion miles of permanent injury-free driving, regulators will be forced to take action by constituents and insurance will skyrocket for cars without the tech. But this is 3+ years after the first commercial robotaxi ride...but that is looking more and more likely to happen this year if we continue to see builds every 2 weeks.

I wanna let folks in this thread know that I threw everything at 12.3 and ZERO safety DEs over 50 miles of urban traffic. If this climbs to 500 we are most likely already on the march of 9s.

 

[DarkandStormy]

Tesla reported ~97k vehicles in inventory at 12/31/23.

Per the 6m tweet, they produced ~440k in Q1 2024, but their displayed inventory is ~90% higher than it was at the end of 4Q23.

If their inventory build grew by 20% (~20k vehicles), then Q1 sales would be the ~440k - ~20k for ~420k).

Of course, the publicly available inventory tracking sites do not capture all inventory, so it's impossible to know.

 

[jschwefel]

I'd suggest folks thinking "RT NEXT WEEK" spend some time reading the threads from users in the FSD forum- you'll see plenty of examples from folks, including the ones who agree 12.3 is a significant improvement, of it still doing outright and unsafe stuff. Myself included.

I think the free trial might well get some extra subscriptions, but the car is still, very very obviously once you've used it for a little while, not ready to do any driving without a human supervising as well as being ready to always fully take over.

As to car carriers, given the current speed of self parking I expect that would actually backlog the logistics lot compared to having a human do it right now... though Tesla claims self park will be getting faster in the future.

And given you still need humans to lock down the car on the carrier it's not like you'd even be saving labor cost.

Yeah, I’m not thinking this is FSD, but literally another NN trained specifically for those tasks. Every car Tesla delivers could be used to train the NN for specific Tesla pre-delivery tasks.

 

[wtlloyd]

Tesla announced that the 6Mth car is produced. Could this be used to estimate the Q1 delivery number?

It's always helpful the read the thread before asking questions/posting news.

*Edit: hung by my own petard, Krugerrand already answered it.

 

[Musskiah]

But this is 3+ years after the first commercial robotaxi ride...but that is looking more and more likely to happen this year if we continue to see builds every 2 weeks.

@Discoducky Hold the phone, did I read this correctly? This year? I agree FSD will march through 9s in a jiffy now that it is not compute constrained. However, the skeptics would point to "pending regulatory approval" and all that jazz. Could you expound upon your thoughts a bit as to how we get around these hurdles in order that the first commercial robotaxi ride occurs in 2024? Thank you for all your FSD insight recently.

 

[traxila]

Well, if that happens, Tesla has already won not only the majority of US vehicle market, but assuredly Europe as well and that sounds like a win. And I do expect that after the first billion miles of permanent injury-free driving, regulators will be forced to take action by constituents and insurance will skyrocket for cars without the tech. But this is 3+ years after the first commercial robotaxi ride...but that is looking more and more likely to happen this year if we continue to see builds every 2 weeks.

I wanna let folks in this thread know that I threw everything at 12.3 and ZERO safety DEs over 50 miles of urban traffic. If this climbs to 500 we are most likely already on the march of 9s.

I am impressed with 12.3 for certain. But I guess NYC is a different animal. Just yesterday disengaged after a few miles due to lane selection that did not exist after having overlooked aggressive and stupid braking for a merge.

Drove the X the rest of the way. Much better. Great car, BTW! lol.

Eagerly awaiting further refinements. Makes sense that crappy urban environments would be the last figured out…

 

[Skryll]

Ok

Yesterday I volunteered at a delivery center giving 12.3.1 demo drives in a 2024 Model S to people picking up new vehicles. People are very impressed (even I was impressed as it was my first experience with vision-only auto park). Car worked very safely on every drive and parking session, no panic/unsafe moments in a 8 min round-trip drive in moderate traffic ranging from 25 mph-45 mph streets and parking lots with one round-about.

I was very excited about getting v12.3 on my 2018 model 3 rwd, and it does a lot of things better but far from what you describe.

Dips are detected now instead of plowed through, it respects the 15mph sign in my residential street instead of barreling down at 25, turns are a bit smoother and faster at stop signs but still need accelerator to be reasonable.

Yesterday after entering the freeway it failed to get out of the lane that led to the next exit and only after exiting it realized its mistake and was attempting to unsafely cross back over through the truck weigh station. I stopped it and completed the exit and reentered on the next one.

A blind left turn it tried to take was too risky for me, not sure if it would have worked without crash so I aborted it:

This is 2023.44.30.25 beta v12.3 so hopefully I get updates resulting in more like you describe - I also wonder if the less quality cameras make a difference. HW3 supposedly is not the reason, Elon x'ed that HW4 is emulating hw3 at this point still ?

 

[2daMoon]

Lobbyists may insist that this tech is in all cars, and insist Tesla license it at affordable prices.

Even so, Tesla will be raking in the buckaroos due to SAAS with an economical hardware component still having more than adequate margins.

Besides, the "first one's free" aspect of Tesla making it available, saving lives, making people ecstatic over Robotaxi, then having the option to turn it off if legislation isn't appropriate for Tesla makes for a strong bargaining position.
(I don't think that it would ever come to this, but it is good to know that we are in the Cat-bird seat)

 

[Musskiah]

How many Tesla owners are in the US right now? Like 1.5M was it?

Just heard the answer from SMR...about 2M

So if only 20% (400k) of the current fleet subscribes to FSD, perpetual quarterly profit boost is $240M or approx 10% (of Q423 profit) every Q. As the N American fleet grows by 1M cars, even if take rate stays at 20% subs @$200/mo, $120M (5% boost) in additional profit accumulates. Ooooooooohhhhweeee

 

[insaneoctane]

Someone remind me, is the Xiaomi SU7 state sponsored and /or are they losing money on each of these? Seems like China continues to work overtime to actually bring the competition and is trying to buy marketshare...If state sponsored, I expect them to retail for significantly more outside of China, but sure makes profitability inside China challenging !

https://twitter.com/x/status/1774080448806728058