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

[SOULPEDL]

It's hard to keep TSLA down, huh? Volume is quite strong today and there was/is a valid attempt to pull it back. Looks like it could fail, maybe next time. RSI is still 65 (but dropping).

As they say...

"The longer the base..."

 

[Thumper]

Turn off blue tooth on your phone while car washing. Trunk and door handle problem solved.

 

[jkirkwood001]

'Interesting pivot' indeed.

We know Ferrari makes tons of profit per vehicle, but I have to believe this is a smarter long-term play than trying to beat Tesla at the supercar owners market, especially if Rimac can be first within certain markets (before Tesla). I still wouldn't invest in them now, but maybe it's a smarter play than status quo EV development...?

Rimac is shifting from electric supercars to robotaxis

 

[growler23]

"The longer the base..."

MMs: "All your base are belong to us"

Re: Rivian
They might still thread the needle here. They got the runway to finish their Georgia factory, and to hopefully finish reducing costs at their original plant. They might (nay, should!!!) defend their internal software from any outside hands for the next year or two, to keep it on track. They can start laying plans for some sort of awful mind-numbing scale of software integration between their world and Cariad (VW software world) but not implement any of those plans until they are gross margin positive on one or more models.
Maybe they can take the FSD transfer approach for defending the code purity of their existing software (just leave it alone for one more quarter, guys! Just one more!)

 

[Musskiah]

Does anyone here even drink anything other than the Tesla kool-aid?

Me? I prefer sour beer, particularly Rare Barrel or Almanac or The Bruery (from you neck of the woods) - if we are talking Cali sours. Cantillon is my favorite outside the US. Judging by your disposition, I wouldn't be surprised if you preferred sours to kool-aid as well.

 

[SOULPEDL]

Here's my FSD code milestones (NOTE: There are other separate milestones for the app, non-realtime interventions and vehicles logistics)

1. >X # of miles between interventions unlocks supervised FSD <-- We are currently here with 12.3.x
2. >Y % over 1B+ miles of completed, successful, no intervention rides unlocks supervised Robotaxi
3. <Z % over 2B+ miles of remote, non-realtime interventions needed unlocks unsupervised Robotaxi

Love the metrics. I'm guessing #2 is where we are at now with 12.4 (depending on the value of X of course). But my bar is pretty low for this since folks are already pretending just fine. 8/8 (or sooner) seems reasonable here.

I strongly feel a #2.5 for remote Summons from anywhere in NA is an important phase. Having zero occupants gathers unique data on actual non-intervention type methods and without risking the driver. The feature also shows progress toward accepting #3 by the public. When we see them driving unoccupied everyday, people should be more inclined to hop in the back.

 

[Musskiah]

XPeng Founder & CEO, He Xiaopeng, on FSD:

"I am truly impressed by the significant progress FSD has made in just a few months. It offers valuable features and a user experience that we can learn from.

I believe 2025 will be a pivotal ‘ChatGPT moment’ for the fully autonomous driving era, revolutionizing how we interact with technology and fundamentally transforming the way we travel.

In the autonomous driving industry, everyone is learning from each other and iterating quickly to create a better mobility experience for all."

That's a decent endorsement from someone menaingful, eh?

XPeng's founder tested out Tesla FSD in the US to see how it compares to XNGP [Video]

For a couple of years, two of the leading driving systems pushing the limits of automotive autonomy have been Tesla’s...

electrek.co

 

[Doggydogworld]

I need to edit this. The current recall was through June 6th (apparently they stopped production?). So that's 7 weeks. 7,810 / 7 weeks = ~1,115/week, if flat.

It's 9 weeks. April 4 - June 6. I laid this out in my earlier reply to you. You can find the actual notices with applicable manufacturing dates at the NHTSA recall website.

 

[Waiting4M3]

On robotaxi, I think it's more important to show that existing HW3/HW4 cars can and will be part of the network, so Tesla will actually have a fleet size advantage right out of the gate. I think to this end, they will roll out the network before the purpose-built robotaxi is available.

I also think they might follow Waymo's route of having safety drivers 1st. As Elon mentioned that they will use remote operators same as Waymo to extract robotaxis that get into trouble. I think Waymo has a lot of smart people, their mistake on lidar vs vision is not born of stupidity, but probably victim of being too early to the game 10 years ago and didn't have the necessary compute to do vision only, then fell victim to corporate inertia and their early success when their approach got too entrenched. On deployment, I think Waymo is doing the right things. I went to SF for the 1st time in ~ a year last month, and was shocked at how many Waymos there are running without a driver, and how comfortable people are with them. It's definitely working imo. Tesla can follow, but progress much quicker due to their overwhelming fleet size.

 

[Discoducky]

In Jordan's new video on Etherloop he gets a lot right, but also misuses some standard terms (which is fine and maybe easily overlooked or benign)

Very short summary on the Etherloop: Generalized redundant hardware with guaranteed virtualized specific and unique necessary dynamically allocated resources.

Another way to say it: Two identical Etherloop compute hubs which are, in effect, like data centers or virtualized compute clusters.

Or another way: Take all the uniqueness and differences of the controller hardware and instantiate it into software or simplify the design to work with existing software.

Or think about it like an equation: Add up all the compute of all the controllers, all the memory, networking and power needs, redundancy, control loop, stability and reboot cycle time requirements. Take the lowest common denominator of all those and build one hub to operate and meet those requirements.

This looks like a data center, with a server, that instantiates a set of virtualized resources, managed by a trusted hypervisor, that boots into dedicated resources, with a dedicated operating system, which can execute commands with guaranteed results.

Why do this? It enables unboxed, saves a ton of money on hardware, cuts down on weight, manufacturing time, allows for all issues to be software issues thus allowing for faster break/fix cycles.

Why hasn't this been done before? You'd need to first have the vision of doing this, hire the right people and put a crazy amount of faith that it could be pulled off as it is a boat load of software engineering at the lowest levels of compute.

 

[Captkerosene]

Positove voices get shut down.
....negative are.just freedom of.speech. the world in 2024.

Content matters not.

A little more respect for the truth is a good thing.

 

[Musskiah]

Positove voices get shut down.
....negative are.just freedom of.speech. the world in 2024.

Content matters not.

The thumbs down to this comment are so revealing.

 

[Usain]

Here's my mental experiment on why it not only makes sense for Tesla to show the vehicle, demonstrate how the experience will work, but also start (at least announce a start date for the existing fleet) supervised Robotaxi's.

Why? Hypothetically, lets say that FSD had hit the milestones for months of driving without an intervention. Then, what would be the next step? Well, you'd want to start doing rideshare and exercising the app experience, pick-up/drop-off experience, wait time management, route and re-route logistics, pricing/payout algorithm and definitely start with drivers/owners behind the wheel. NOT just jump to driverless. Removing the driver wouldn't even come next. You'd want to then let the remote operator, algorithm or whatever non-realtime system is used as back-up then get exercised while a human is behind the wheel. Once that is solid, with all the other experiential stuff and you have a very high confidence in successful rides (like 1000 to 1) vs rides which required a non-real-time intervention.....then the next step is removing the human driver.

Here's my FSD code milestones (NOTE: There are other separate milestones for the app, non-realtime interventions and vehicles logistics)

1. >X # of miles between interventions unlocks supervised FSD <-- We are currently here with 12.3.x
2. >Y % over 1B+ miles of completed, successful, no intervention rides unlocks supervised Robotaxi
3. <Z % over 2B+ miles of remote, non-realtime interventions needed unlocks unsupervised Robotaxi

This has been my thinking exactly.

8/8 will be more than just a vehicle reveal. I believe there will be a concrete announcement about startup of the supervised robotaxi service.

 

[aubreymcfato]

Does anyone here even drink anything other than the Tesla kool-aid? It's so strange that everything anyone else is doing is always majorly bad when to your 1st point, the whole gen2 R1T/R1S updates were focused a lot on true cost-saving measures that you said weren't even considered. Maybe watch/read about the gen2 R1 platform before shooting off misleading info.

Rivian unveils next-gen R1 EVs to cut cost without affecting driver experience

Rivian has unveiled its next-generation R1S and R1T electric vehicles to cut cost without affecting the driver experience negatively. In...

electrek.co

Lots of changes:

Review: 2025 Rivian R1T and R1S gain efficiency, comfort, and a lifeline

Rivian has made its R1 lineup lighter, more efficient, and less complicated to build—and better to drive.

www.greencarreports.com

This is not a Rivian forum, but bottom line is Rivian simply needed $$ to be able to get to releasing R2/R3 and this was the deal they chose. New gen 2 trucks can go up to 420 miles of range vs. 300s for Cybertruck. Yeah, I get stock investors here are biased and wants to see the competition all die, but total head in the sand with your comments.

Tesla margins were high because they were the only EV game in town (outside of Bolt/Leaf). Their net margins were 5.5% in Q1 (used to be close to 20%). which is legacy auto numbers. Remember the $68k MY? Tesla gross margins is like legacy auto makers (Ford has like $14k gross margins on the F150 I've post/found before).

Please.
Being this a Tesla forum, Rivian is always touche with silky gloves: I don't wander in Rivian forums, but I wouldn't expect the same courtesy. Being from Europe, I know Rivian cars just from what I read here, and I know they are nice, beautiful cars with great interiors and software.
What I also know is that they always need money, they didn't give Amazon the millions vans they asked, and know also VW is involved.
I think I wrote again in this thread that I hope they succeed, but proof is always in the pudding: in 2023, they produced 57232 vehicles and delivered 50122. Tesla instead produced 495000 vehicles and delivered over 484000 vehicles. That is almost 10x.
As my nonna would say, Rivian need to "eat a lot of bread" to be able to seat at the adult's table with Tesla.

And the last paragraph is even more confusing: are you really mixing ICE profits with EV profits. Where are the other "pure EV companies" profits? Rivian is the second best game in town in US, and we are talking about a "lifeline" from VW.

 

[jerry33]

This is the clearest indication possible that VW haven't figured out a solution to their software problems. They must have recognized that their software team is incorrigible. I hope that the joint venture is firewalled and able to develop from Rivian's coding without the lethargy and numerous duchys established within VW's program.

They still have the problem that a lot of items that require program access are outsourced and every outsource supplier has a different program interface. I don't see how Rivian can help this--although I'm sure it looks good on a presentation slide.

 

[Knightshade]

That's a decent endorsement from someone meaningful, eh?

Is it? FWIW I agree with you- but when folks previously mentioned here Xpeng offers "works anywhere in the country" L2 city streets software- the country being China rather than the US, in reply to others claiming nobody has anything remotely close to FSD anywhere, the response was largely to dismiss Xpeng as meaningless.

What I also know is that they always need money, they didn't give Amazon the millions vans they asked,

TBF, Tesla was founded in 2003, and was still doing cap raises by selling stock as late as December 2020. Rivian was founded 2009, so they've got until end of 2026 to stop raising cap via stock sales to be even with Tesla in that regard.

Also the amazon thing was 100k vans, not 1M-- and by 2030, not this year. And they're still working toward that goal- only change was the van sales are no longer exclusive (ie they previously were contractually bound to ONLY sell em to Amazon, that's no longer true)
See here:

Rivian to end exclusivity with Amazon, allow other companies to buy its electric vans

Electric vehicle maker Rivian says it will allow other companies to purchase its delivery vans, ending its exclusive agreement with Amazon.

apnews.com

Rivian reiterated its pledge to help Amazon put 100,000 electric vehicles on the road by 2030. Amazon says it’s rolled out more than 10,000 so far.

I think I wrote again in this thread that I hope they succeed, but proof is always in the pudding: in 2023, they produced 57232 vehicles and delivered 50122. Tesla instead produced 495000 vehicles and delivered over 484000 vehicles. That is almost 10x.

Sure. But that's Rivian just over 2 years from first production of a mass produced vehicle (Sept 2021 was when that began)... Versus Tesla 11 years after first production of a mass produced vehicle (June 2012 for Model S). FWIW to reach 50k deliveries Tesla didn't get there until 2015... Rivian got there a year faster than Tesla did from start of production.... (In 2014, 2.5 years after production start, so slightly MORE time than Rivian had end of 2023, Tesla only delivered 31,655 vehicles vs. Rivians 50122).


FWIW there's lots of OTHER ways Rivian is fundamentally poor compared to Tesla (even early production Tesla)-- but not those ways.

 

[sunwarriors]

Dude what are you talking about?

I didn’t disparage Rivian at all. Nor did I say I wanted Rivian to die.

Just said that one more big hand in their pot is not going to be a net positive for them. Which I think most people would agree with. But that triggered you for some reason…

Rivian developers now have a big distraction. Rather than continuing to make improvements for Rivian vehicles, they now have to support a whole bunch of other vehicles from VW.

And are you seriously comparing ICE margins to Tesla’s EV margins? Lol. The challenge has been in making EVs profitable. ICEs have been profitable for over a hundred years.

Your 1st point said they didn't look at cost savings at all which was totally wrong here which makes it look like you had no clue really other than spouting incorrect info (positive FUD). Similar to how posters don't like incorrect information posted, you were posting totally incorrect info so I called it out.


As I state again, Rivian needed $$. There are articles which state Ford, even though a big early partner of Rivian didn't step up because Ford has a massive truck business which competes with Rivian. VW, especially in the US, has none of that (that I know of). Similar to folks thinking legacy OEMs would jump to license Tesla FSD, there are very large competitive reasons (pay your competitors directly) that is against most corporate cultures to support/license FSD, not to mention, they are all developing their own driver assist/have their own AI divisions.

Similar to every single company out there developing their own AI chips, no one wants to fund/support a competitor to kill off your own business. Dan Ives, Wedbush uber TSLA bull even felt this was "...a game changer..." for Rivian. Read it for you will, but kool-aid drinking investors choose to see what they want, especially when a big TSLA bull thinks this is a positive for RIVN as a company due to their financial issues.

ICE margins at the gross level in the 15-20%, net was what TSLA had justifying their higher valuation (it used to be close to 20%), that's gone now and it's RT/AI/Robotics, hence the Elon shift/accelerated change.

 

[jerry33]

Is it? FWIW I agree with you- but when folks previously mentioned here Xpeng offers "works anywhere in the country" L2 city streets software- the country being China rather than the US, in reply to others claiming nobody has anything remotely close to FSD anywhere, the response was largely to dismiss Xpeng as meaningless.




TBF, Tesla was founded in 2003, and was still doing cap raises by selling stock as late as December 2020. Rivian was founded 2009, so they've got until end of 2026 to stop raising cap via stock sales to be even with Tesla in that regard.

Also the amazon thing was 100k vans, not 1M-- and by 2030, not this year. And they're still working toward that goal- only change was the van sales are no longer exclusive (ie they previously were contractually bound to ONLY sell em to Amazon, that's no longer true)
See here:

Rivian to end exclusivity with Amazon, allow other companies to buy its electric vans

Electric vehicle maker Rivian says it will allow other companies to purchase its delivery vans, ending its exclusive agreement with Amazon.

apnews.com

Sure. But that's Rivian just over 2 years from first production of a mass produced vehicle (Sept 2021 was when that began)... Versus Tesla 11 years after first production of a mass produced vehicle (June 2012 for Model S). FWIW to reach 50k deliveries Tesla didn't get there until 2015... Rivian got there a year faster than Tesla did from start of production.... (In 2014, 2.5 years after production start, so slightly MORE time than Rivian had end of 2023, Tesla only delivered 31,655 vehicles vs. Rivians 50122).


FWIW there's lots of OTHER ways Rivian is fundamentally poor compared to Tesla (even early production Tesla)-- but not those ways.

I believe Rivian's big problem now will be to keep VW's managers and corporate culture out. Looks as if they've done an okay job with the others.

 

[Nathan]

Once of these places is NE Ohio. I drive 100+ miles per day on average, mostly city streets, 99%+ of the time on FSD. Since 12.3.x, I have gone weeks without an intervention. I have a 2021 MX (pre-refresh) with no internal camera and a 2022 MY. Both HW 3. I know several Tesla owners in the area with similar experince. Anecdotally, here in Ohio, FSD seems ready for professional safety robo-copilots.

I have been driving Autopilot since 2016 and FSD since early inception (I had a 99 safety score for the original rollout), so it is possible I am much more trusting than the average user of FSD. Nonetheless, I am accident-free with FSD; in fact, as I have shared many times here, FSD has saved me from at least four bad accidents where I would not have been at fault (car ran a red light in a white-out storm preventing a driver-side collision, box truck tried to merge into me from my blind spot on on the highway at 2am on the way to Maine, etc).

The areas requiring interventions have generally been construction zones with signs that FSD cannot read...yet. No doubt this limitation is very temporary.

Driving in North Carolina near Asheville and having quite a few interventions.
Still going too wide on some windy rides around corners . Other times it's great.
I've been having lots of issues with speed.
choices .Sometimes it goes under the max speed setting on many roads for no apparent reason
just staying at 49 in 45 mph zone or going 37mph in 45 again for no reason
Couple of times when car in front of me was signaling right my car was heading for the car to some degree and not anticipating and moving into left lane when it was entirely clear . Other situations too
Though I will say on the very curvy ride up to my mountain property (did you hear that Krugerand ?) it was doing largely better than it ever has . So there is a long way to go from my perspective in my driving in North Carolina and Brooklyn, NY. So FSD still has a lot to learn.
I've had FSD from 2018.

 

[Tony73]

TBF, Tesla was founded in 2003, and was still doing cap raises by selling stock as late as December 2020. Rivian was founded 2009, so they've got until end of 2026 to stop raising cap via stock sales to be even with Tesla in that regard.

I don’t believe Tesla has ever had a down fundraising round as private or public company. Rivian went public at around $70. It’s now raising funds at $10. That is a huge disaster.