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

[Big Time]

LOL "A TREMENDOUS NEED FOR ASSWIPING" I love this man

 

[Bet TSLA]

Analyst: If you're not doing a 25k car and you said you're thinking 250k cybertrucks a year- how do you get to 3.2 million or more vehicles in 2024 per >50% YoY guidance?

Elon: YOU DO NOT UNDERSTAND FSD!


...what?

I think it's clear that Elon sees that anybody who can afford a $25K car, but not a Model 3, would find it an obvious choice to just use a robotaxi. So FSD means the market moves to robotaxi fleets.

 

[RobDickinson]

If FSD is solved, no 25K car

No need for a $25k car when a $20k bot has replaced your job!

 

[hacer]

For all the many, many quarterly calls during which Mr Musk has spoken glowingly of FSD.....he is now doubling...tripling...quadrupling down in putting his reputation and believability on the line with its real, $ impact on Tesla in CY 2022.

Nobody believed his time lines after the first couple of years. His reputation on that question is already toast. So lets hope nobody seriously believes it will impact TSLA CY 2022.

 

[tivoboy]

okay, Bonus points for swearing on an investor call! Who had f.. g on their bingo card?

Update. Double word score points if you also had “a..s wiping” on your bingo card.. I think someone paid them real $$ to somehow introduce these words or phrases into this call. It’s almost too obvious.

 

[mhan00]

I would assume that Tesla thinks they have the demand and can ramp S/3/X/Yproduction to meet their 3.2 million projection without a 25k car.

 

[Knightshade]

Confirmed no 4680 iron cells- which I think debunks an idea somebody was pushing in here a while back

 

[OrthoSurg]

With the 2.85 EPS and 50% Guidance, I think the AH should be at least what MSFT scored yesterday

 

[Discoducky]

My interpretation of what Elon is saying is that they do NOT need to make a $25k car at this point as FSD will arrive significantly before the $25k car would be profitable.

In other words, FSD is going to unlock the robotaxi fleet and will fundamentally change how people move about the world at somepoint (as taking a subsidized public bus will be more expensive).

In other words, demand for all urban transportation needs will be significantly reduced due to robotaxi.

And I agree with folks, that Elon is now as confident as ever in FSD being solved this year.

So, a timeline is sounding like:

end of 2022 - FSD is safer than a human (Elon would be personally shocked if not achieved)
end of 2023 - Level 4 (my interpretation)
end of 2024 - Level 5 (my interpretation)

 

[2daMoon]

LOL "A TREMENDOUS NEED FOR ASSWIPING" I love this man

... then they talk about diarrhea,

or, maybe that was die area.

 

[henchman24]

The basics of what Elon was saying is people will pay 50-70k for a Y if they can farm it out for half that cost and then use the other half (clearly Uber will be all over as many as they can buy). So they can just produce a few million Ys and everybody will buy them. They don't need to add a ton of models nor go cheaper if they can just make the cost of ownership factors cheaper. Along with that Tesla will have their won fleet that people will just pay to use instead of buying their own car.

 

[Boomer19]

id prefer the main narrative be unadulterated explosive growth, built upon their fundamental strengths (engineering, ops, logistics)

and, bonus, 4680 in the products right now, since that is the future foundation of products.

- let fsd/robotaxi and optimus be the whip cream on top, a year or whatever time is needed down the road

 

[mrmage]

Dojo will "do something" by this summer, but maybe will be better than existing GPU clusters next year.

Claims Dojo not needed for FSD.... and they're not sure yet it'll succeed (Dojo that is), but they think it will.

I was surprised by this too. The GPU approach seems to be making enough progress that Dojo may not be needed. Ultimately, Elon says it's the engineers who will decide if Dojo is better and worth switching to. That being the case, Dojo does not sound like an obvious win at this point.

 

[Bet TSLA]

end of 2022 - FSD is safer than a human (Elon would be personally shocked if not achieved)

When was the first end of year when Elon would be shocked if FSD wasn't better than human drivers? Was it 2019 or 2020? He has no credibility whatsoever on this subject any more. He's playing Lucy and the football to our Charlie Brown.

Predicting software is really hard. There are *no* first principles in software, which is probably why Elon is no good at it. But regardless, I wish he would just stop talking about it when all he can do is make his reputation even worse.

 

[Beltsbear]

Which means we should see energy ramp to be significant this year, hence the megapack factory.

Not necessarily. Megapacks need chips too. Lots of voltage switching and conversion chips.

 

[AudubonB]

I do not know if it is a Google, or YouTube, or Tesla I.R. team that is in charge of the subtitles for the conference call video but -

for me, a magnificent confidence-builder in Tesla's AI program, whether Dojo or not, would be to have IT in charge of creating those subtitles. If it can 99.xxxxx% correctly write what esp. Mr Musk is saying, then driving through Mumbai's rush hour or my typical herd of winter-zombie caribou will be a piece of cake.

Gauntlet thrown down! Respond to that, Tesla!

 

[Big Time]

The basics of what Elon was saying is people will pay 50-70k for a Y if they can farm it out for half that cost and then use the other half (clearly Uber will be all over as many as they can buy). So they can just produce a few million Ys and everybody will buy them. They don't need to add a ton of models nor go cheaper if they can just make the cost of ownership factors cheaper. Along with that Tesla will have their won fleet that people will just pay to use instead of buying their own car.

For a while I had this notion in my head like "Who would rent out a 50K car to strangers" but the further that Tesla implements its own insurance, as well as the quality of the cameras inside and outside the car, if some knucklehead wants to $%$# in the backseat or a drunkard vomits, than Tesla just pays the bill, and that user is discredited. Similar to what air BnB does but with actual monitoring devices. Pretty amazing

 

[Knightshade]

The basics of what Elon was saying is people will pay 50-70k for a Y if they can farm it out for half that cost and then use the other half (clearly Uber will be all over as many as they can buy). So they can just produce a few million Ys and everybody will buy them. They don't need to add a ton of models nor go cheaper if they can just make the cost of ownership factors cheaper. Along with that Tesla will have their won fleet that people will just pay to use instead of buying their own car.

I'm not sure how they then get to 20 million cars a year before 2030, as previously announced as a target.... If you're replacing every 5 cars with 1 robotaxi--- there's not nearly 100 million cars sold NOW per year.... if you were replacing all new cars 5:1 with RTs you'd only have what like 10-15 million TOTAL sold in a year? (for the entire vehicle market combined)


One thing I kinda wish someone had asked about... Elon previously mentioned a next-gen sensor suite for FSD (upgraded cameras and HW4 computer at least) and suggested it would be likely be introduced in Cybertruck in mid-late 2022. Now that we know no CT this year- is the sensor suite likewise pushed back till 2023 at least?

 

[Jeeps17]

With the 2.85 EPS and 50% Guidance, I think the AH should be at least what MSFT scored yesterday

I agree, but this being TSLA, logic rarely applies for very long (except over multi-year timeframes).
Q4 delivery beat effect lasting 12 seconds (I jest?) is the latest example.

Still HODLing my shares since 2012, irrespective of short-term noise.

 

[Discoducky]

When was the first end of year when Elon would be shocked if FSD wasn't better than human drivers? Was it 2019 or 2020? He has no credibility whatsoever on this subject any more. He's playing Lucy and the football to our Charlie Brown.

I'd agree, that Elon has been burnt pretty badly with these FSD predictions.

And from my experience developing large scale billion+ parameter AI models, training them, putting them into production and running very large scale inference...this is hard. So many layers that can show many false positives for no local maxima.