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

[Mengy]

I think 8 million is pie in the sky. Maybe 3 or 4 million. And that’s only if they expand quickly into untapped markets like South America.

The ticket will be finding markets where they don’t associate Tesla with Elon. That’s just the way it is now. At some point even tesla will have to recognize that.

I think tesla could still be a good and profitable company at those levels. Just not the industry behemoth everyone thought it would be.

Jmho.

I believe Tesla will be way above 4 million autos produced per year by 2030. They are just below 2 mil now, once interest rates drop Tesla can ramp the lines back up and hit 2 - 2.5 mil easily. Add 250K CT's per year, 50K Semis, 1 or 2 million of whatever the economy 3/Y's might be, and then 3-4 million RT's per year (with Giga Mexico and at least one more Giga by then) by 2030 and you're at 8 mil run rate. Maybe throw in 10K Roadsters per year if it ever actually gets made.

I feel 8 mil/yr by 2030 is a conservative yet realistic number. I think the only way it gets higher than 8 mil is if Tesla ramps RT production super aggressively, which is a possibility IMHO.

Plus the actual Model 2, the compact consumer unboxed Tesla, might still get made down the road.

 

[Singuy]

That his videos misrepresent the facts and can't be trusted?

Sure, but we've known that for years- hardly the first time he was dishonest in a video about self driving--- and why repost such trash here?




How?

NOTHING in Omars video is about what mercedes said about anything except the parts where the other dude in the car keeps pointing out they're using it in conditions Mercedes says it's not meant to work but Omar keeps counting as "interventions" anyway.


If I pushed out a video where I only enabled dumb cruise control on a car-- then counted as "interventions" every time I had to hit the brake to not hit a car in front of me- would that seem a reasonable use of anyone's time or an attempt at any sort of honest point? That's what he did here.

Notice how -0 percent- of the video is on an actual highway? Because he's intentionally doing all he can to operate the system where it's not intended and make it look bad.... then he doubles down when the other dude tries to make the reasonable point it's ONLY FOR HIGHWAYS and Omar tries to tell him "Any road is a highway if you're doing 45".... what?




But the system IS LEVEL 3.

The fact your refuse to learn what that even means is not Mercedes problem- it's yours.





What CR said or didn't say has nothing to do with OMAR being dishonest in the video my dude.


Your argument has now distilled down to "I don't like what others said about Tesla, so I'm fine if other pro-Tesla people say dishonest things in response"

That's a garbage argument.

Why are you making it?

What? It's not rocket science to understand that if Mercedes try to position or market themselves as a leader in autonomy by certifying a useless system as L3, then they better bring it on and test themselves against what most people thinks is the best commercial autonomy system.

Look if that L3 system wasn't so useless, and it actually moved autonomous tech forward, I don't think people would have a problem with it.

Also no one here said their system isn't L3 so don't be putting words in my mouth. Most of my argument is about L3 being a garbage SAE level to begin with (and all of SAE levels are garbage when used naturally by people to compare capabilities when it's not even meant to be used as such).

 

[Rarity]

Haven't seen this mentioned here. I saw it first on Solving Modey Problem YT-channel. In nutshell, Hyundai is retreating from robotaxi development.

"Hyundai Motor Group's joint venture Motional announced staff reductions and a two-year delay to its driverless taxi service plans last week, following the group's 1.29 trillion won ($941 million) injection into the startup earlier this month, reflecting growing industry pessimism about the viability of self-driving cars.

Hyundai established Motional with Aptiv in the US in 2020 and launched its Ioniq 5-based robo-taxi service with Uber and Lyft last year. However, the rollout of the second-generation, fully autonomous service based on the Hyundai Ioniq 5 has now been postponed to 2026 -- two years later than initially planned.

This delay includes halting current operations such as taxi services in Las Vegas and delivery services for Uber Eats in Santa Monica, all of which currently require a human safety operator. According to the company, the services have completed more than 100,000 rides in Las Vegas and food deliveries in Los Angeles.

"Hyundai Motor Group's joint venture Motional announced staff reductions and a two-year delay to its driverless taxi service plans last week, following the group's 1.29 trillion won ($941 million) injection into the startup earlier this month, reflecting growing industry pessimism about the viability of self-driving cars.

Hyundai established Motional with Aptiv in the US in 2020 and launched its Ioniq 5-based robo-taxi service with Uber and Lyft last year. However, the rollout of the second-generation, fully autonomous service based on the Hyundai Ioniq 5 has now been postponed to 2026 -- two years later than initially planned.

This delay includes halting current operations such as taxi services in Las Vegas and delivery services for Uber Eats in Santa Monica, all of which currently require a human safety operator. According to the company, the services have completed more than 100,000 rides in Las Vegas and food deliveries in Los Angeles."

Motional delays self-driving service launch to 2026 following Hyundai investment

Hyundai Motor Group's joint venture Motional announced staff reductions and a two-year delay to its driverless taxi service plans last week, following the group's 1.29 trillion won ($941 million) injection into the startup earlier this month, reflecting growing industry pessimism about the...

m.koreaherald.com

Assuming that this retreat is true, it would seem to indicate a retreat at Boston Dynamics as well. I am not sure that Boston Dynamics was ever a potential true competitor to Optimus, but it is an interesting measure of appetite for AI investing nonetheless.

 

[hobbes]

April 2024 was just <200 cars shy from April 2023.
Some countries are stepping up, but the big 3 are declining: Germany, France and UK are down roughly 12k cars in the first 4 months of 2024, YoY. This means that it's like we lost a whole quarter of UK or France. So there are reasons to be concerned.

A lot of cars were bought in Germany but resold outside the country, likely to take advantage of the Umweltbonus (which is now gone). I had heard a lot of that over the years but recently there was a report by the German Kraftfahrtbundesamt (DMV) on the fleet changes in 2023 which makes it all too obvious: while we saw sales of 15,865 Model 3 and 45,818 Y in 2023, the fleet increased by only 7,446 Model 3 and 36,707 Y. So it looks like 50% of all Model 3 and 20% of all Y sold are not registered in Germany any more!

So to see the whole picture, it is important to look at the European market as a whole. If we see sales drop in Germany more than in other European countries, the above could explain a lot of that.

Fun facts: Model Y fleet more than doubled in 2023, Model 3 was +10%. But there are now more 3s (79,690) registered than Ys (71,774) as of Jan 1, 2024.

@Troy

 

[rentierparasit]

I believe Tesla will be way above 4 million autos produced per year by 2030. They are just below 2 mil now, once interest rates drop Tesla can ramp the lines back up and hit 2 - 2.5 mil easily. Add 250K CT's per year, 50K Semis, 1 or 2 million of whatever the economy 3/Y's might be, and then 3-4 million RT's per year (with Giga Mexico and at least one more Giga by then) by 2030 and you're at 8 mil run rate. Maybe throw in 10K Roadsters per year if it ever actually gets made.

I feel 8 mil/yr by 2030 is a conservative yet realistic number. I think the only way it gets higher than 8 mil is if Tesla ramps RT production super aggressively, which is a possibility IMHO.

Plus the actual Model 2, the compact consumer unboxed Tesla, might still get made down the road.

Hi, Mengy --

> and then 3-4 million RT's per year

Right, so you're also at 4-5MM ex-Robotaxi. This may be hair-splitting, but I'd attribute any RT sales to software/FSD rather than auto.

Yours,
RP

 

[Rarity]

They will be out in 10 cities by the end of next year and that is what they service with the 50k vehicles they bought from Geely. They are basically telling you exactly what they'll do. It's not a secret. They think 50k vehicles covers 10 cities more or less. That's about what Uber has working those cities. SO.....so....Waymo plans to take all the Uber rides. They can too. The cost of the vehicle is noise in the wind. Literally meaningless. Million miles on $50k or $25k vehicles ...it just doesn't matter. Google will make it all back on ads and more.

This view seems to ignore the unit economics, which makes the cost of the robotaxi extremely important. The BG2 podcast details this. This was linked upthread. See discussion starting at 41:30, with the unit economics piece at 58:00.

 

[insaneoctane]

I'll suggest that the Mercedes vs Tesla, L3 vs L2 move to its own thread after I get the last word

I have NOT read the CR article and I have NOT watched Omar's video...but, listening to the bicker here (it's hard to avoid), it sounds like Mercedes out-foxed Tesla with marketing their limited product. It's not clear to me that if you narrowly compare the 2 products within Mercedes required and limited use cases, which is better. Maybe that's what CR did? It might be infuriating and arguably dishonest to claim superiority using such a limited and narrow capability, but that's marketing! Mercedes is taking a bigger risk (ie liability) for their claims if it can't perform properly within the limited restrictions? Tesla could choose to do this, but they choose not to.

 

[Musskiah]

Mercedes​

Required operating conditions​

Building over a century of trust in drivers across the globe begins and ends with safety. DRIVE PILOT is ready to chauffeur you under conditions that help ensure a secure ride. Conditions include:

• Clear lane markings on approved freeways
• Moderate to heavy traffic with speeds under 40 MPH
• Daytime lighting and clear weather
• Driver visible by camera located above driver's display
• 
  There is no construction zone present.

The 1st oxymoron is to be driving on highways at speeds under 40MPH. It got me So just when your commute is stuck.
the fact that they advertise for 40MPH and not on city roads at that speed means it's a piece of crap. (there I said it). At least waymo sticks to city streets at these speeds.
What happened to German Engineering and the Autobahn speeds
How is this L3 I don't know? Don't point me to the architecture docs ( , I have read them )

Under these conditions Tesla might be L5 +

Street/SPACs definitely scheming to creating the Nikola, Lordstown of FSDs

@nativewolf and @sunwarriors and @Doggydogworld don't get it even when you spell it out for them. Perhaps you can make it even more clear than this?

 

[FreqFlyer]

I believe Tesla will be way above 4 million autos produced per year by 2030. They are just below 2 mil now, once interest rates drop Tesla can ramp the lines back up and hit 2 - 2.5 mil easily. Add 250K CT's per year, 50K Semis, 1 or 2 million of whatever the economy 3/Y's might be, and then 3-4 million RT's per year (with Giga Mexico and at least one more Giga by then) by 2030 and you're at 8 mil run rate. Maybe throw in 10K Roadsters per year if it ever actually gets made.

I feel 8 mil/yr by 2030 is a conservative yet realistic number. I think the only way it gets higher than 8 mil is if Tesla ramps RT production super aggressively, which is a possibility IMHO.

Plus the actual Model 2, the compact consumer unboxed Tesla, might still get made down the road.

I think we put too much emphasis on how the Model 2 will be built. A lower-priced product built on lines already amortized is a smart move. Especially if the 48V architecture moves to the best-selling car, that drops costs significantly, and not just the RT, and CT.

 

[Knightshade]

@Drax7 did you miss the mods admonishment against spamming a users posts with the Disagree button in lieu of substantive explanation of your disagreement?

The one thing Tesla has that nobody else has, is the freedom and confidence to say "here is your one month free trial. Drive using FSD anywhere, any time, in any weather, and you decide how reliable it is."

Thats not a phrase a mercedes salesperson will be using with any confidence.

Except Tesla says it with the condition that a human must always be paying attention to the road.

According to some folks in here if the system is ever CONDITIONAL it's not a real system. But I'm sure they only mean when it's not Tesla they're talking about.

Thats not a phrase a mercedes salesperson will be using with any confidence.

They won't be using that phrase at all since they understand their system is explicitly not intended for that. Heck- the owner of the Mercedes IN THE VIDEO understands that-- and pointed it out to Omar several times.... Omar first ignores him, then lies about what a "highway" is, to deflect from the fact he's making a dishonest video, not for the first time.

How is this L3 I don't know? Don't point me to the architecture docs ( , I have read them )

Have you?

Because it's L3 since it can perform the entire DDT within it's ODD, and a DDT fallback to allow a reasonable amount of time for a human to take over when needed beyond that. That's it. That's the basic definition of L3.

Under these conditions Tesla might be L5 +

But it factually is not- because it doesn't have a complete OEDR capable of performing the entire DDT-- the human must fill that role. It also has no DDT fallback (though that'd be much easier to add than completing the OEDR)

What? It's not rocket science to understand that if Mercedes try to position or market themselves as a leader in autonomy by certifying a useless system as L3, then they better bring it on and test themselves against what most people thinks is the best commercial autonomy system.

It's not rocket science, it's nonsense circular argument.

And has nothing to do with the video being outright dishonest in multiple ways.


"It's ok for Omar to lie a bunch because Mercedes factually said their system is level 3" is a dumb argument.

Why are you making a dumb argument?

It's not clear to me that if you narrowly compare the 2 products within Mercedes required and limited use cases, which is better. Maybe that's what CR did? It might be infuriating and arguably dishonest to claim superiority using such a limited and narrow capability, but that's marketing! Mercedes is taking a bigger risk (ie liability) for their claims if it can't perform properly within the limited restrictions? Tesla could choose to do this, but they choose not to.

100% agree.

 

[Webeevdrivers]

Let’s hope.

I believe Tesla will be way above 4 million autos produced per year by 2030. They are just below 2 mil now, once interest rates drop Tesla can ramp the lines back up and hit 2 - 2.5 mil easily. Add 250K CT's per year, 50K Semis, 1 or 2 million of whatever the economy 3/Y's might be, and then 3-4 million RT's per year (with Giga Mexico and at least one more Giga by then) by 2030 and you're at 8 mil run rate. Maybe throw in 10K Roadsters per year if it ever actually gets made.

I feel 8 mil/yr by 2030 is a conservative yet realistic number. I think the only way it gets higher than 8 mil is if Tesla ramps RT production super aggressively, which is a possibility IMHO.

Plus the actual Model 2, the compact consumer unboxed Tesla, might still get made down the road.

But I don’t see Giga Mexico happening. Not unless there is a new smaller model. And that I think is at least for now off the table.

Jmho.

 

[GatorMeat]

Tesla's Model 2 reportedly shelved just before completion

Tesla Inc. and SpaceX Chief Executive Officer (CEO) and X Corp. Chairman Elon Musk canceled the production of the electric vehicle (EV) company's Model 2 when it was nearly done, The Information repor...

breakingthenews.net

More news today of the Model 2 cancellation/shelving/insert term here.

This makes absolutely no sense to shelf a new high volume model right before production when you have lots of spare existing capacity. Austin Model Y night shift was canceled to retool the line for the new model. Berlin production has also been throttled. Read the tea leaves

 

[SarahsDad]

Well, at least those who keep worrying about "What will happen after the first autonomous fatality" are now more likely to get their answer, only it may come from an incident with a Mercedes rather than a Tesla.

Or Amazon…

Amazon's Zoox under investigation by NHTSA after two robotaxi crashes

Two Toyota Highlanders equipped with Zoox's automated self-driving technology braked unexpectedly, causing motorcyclists to rear-end the vehicles.

www.cnbc.com

 

[Usain]

The loophole comes from the garbage SAE that doesn't clearly define capabilities and yet assign them levels in which you can have extremely capable autonomous features as L2 and garbage capable systems as L3.

You must not question the SAE. Theirs is holy writ, the infallible word of God.

 

[Todd Burch]

We've seen Tesla slowing auto production, possibly changing the Model 2 into an economy 3/Y variant, dropping auto margins to razor thin levels,

Where on earth do you see Tesla's auto margins at "razor thin"? They're nowhere near razor thin.

As reported in Tesla's Q1 10-Q filing, automotive gross margins were 18.5%, among the best in the industry:

 

[AndrewZ]

People shouldn’t mistake lack of desire/interest for lack of capability. Mercedes is using NVIDIA’s tech for their ADAS, and NVIDIA has implementations that can attempt city driving like FSD — you can find videos on YouTube of Mercedes vehicles doing that, because the two have been partnered for some time now.

No other carmaker wants to put Level 2 City Streets functionality in consumer’s hands and there are no signs that will be changing, Tesla is simply the only company who has any interest in doing it. The other companies are doing their testing and development privately, with professional testers etc, and they won’t release anything to consumers until it was actually fully ready to go with minimal risk.

 

[Mengy]

Where on earth do you see Tesla's auto margins at "razor thin"? They're nowhere near razor thin.

As reported in Tesla's Q1 10-Q filing, automotive gross margins were 18.5%, among the best in the industry:
View attachment 1046751

Yeah, "razor thin" was probably the wrong descriptor for me to use. Much thinner than they used to be would be more accurate.

 

[Doggydogworld]

The trouble with fighting about rival self driving systems using real time video is that its not a single drive that is the problem. I'm sure its perfectly possible to do a 90 minute Mercedes drive that is flawless. That would make for a great stock-pump video.

Yes, 100%. Same is true for Omar's stock pump videos.

The problem is, if you only have a catastrophic crash that kills your children with an autonomous car every thousand miles... thats still probably quite an issue, but it would make for a very tedious 40 hour long video.

The one thing Tesla has that nobody else has, is the freedom and confidence to say "here is your one month free trial. Drive using FSD anywhere, any time, in any weather, and you decide how reliable it is."

Tesla also says if a catastrophic crash kills your children it's entirely your fault. And they'll fight tooth and nail in court regardless of how their "Level 5 just around the corner" system behaved. Mercedes will actually have to pay out. Same with Waymo and others. That's the gigantic gulf between Level 2 and Level 3/4 systems. I take Waymo to task for slow expansion and Mercedes Drive Pilot for a ridiculously limited operating domain, but they stand behind their autonomous technology. Even after a decade of Muskhype and nutty claims on forums like this one, Tesla still doesn't. Maybe some day they will, but until then it's kinda ridiculous to preach about superiority and "10 years ahead" and take potshots at those who are actually dong what Tesla can't.

 

[Doggydogworld]

People shouldn’t mistake lack of desire/interest for lack of capability. Mercedes is using NVIDIA’s tech for their ADAS, and NVIDIA has implementations that can attempt city driving like FSD — you can find videos on YouTube of Mercedes vehicles doing that, because the two have been partnered for some time now.

No other carmaker wants to put Level 2 City Streets functionality in consumer’s hands and there are no signs that will be changing, Tesla is simply the only company who has any interest in doing it. The other companies are doing their testing and development privately, with professional testers etc, and they won’t release anything to consumers until it was actually fully ready to go with minimal risk.

L2 City Streets is actually becoming a thing in China. I agree it's less likely in the US. Europe is focusing on highway L3 and raising the limit from 60 kph to 130 (roughly 37 mph to 80).

 

[nativewolf]

I think that early days in RT service we won't really have any idea what unit economics are like. How much ad revenue can google extract from Waymo? No idea, to me it's clear where Waymo has been headed. They thought vehicle X would get there, they canceled that and bought vehicle Y, it's always about the same number and that, IMO, means they are going after most of Ubers most profitable markets in the USA. The Geely vehicle is a good vehicle, easy to clean, suitable for people with disabilities, and the option for the new super CATL battery.

I would guess it won't really be an issue until several years of operating and fine tuning.

This view seems to ignore the unit economics, which makes the cost of the robotaxi extremely important. The BG2 podcast details this. This was linked upthread. See discussion starting at 41:30, with the unit economics piece at 58:00.