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

[Electroman]

If waymo actually had the same number of driverless miles as Tesla I'd be worried (they are below 0.1% as I recall?)

Tesla has ZERO driverless miles.

 

[Krugerrand]

I mean...how can you type something about the future and make any such assertion about something that has never been done.

The same way I and others knew Tesla would succeed over a decade ago. The same way I and others know Tesla is going to be the largest company on the planet. The same way I knew as a kid I would live on a mountain.

Maybe his numbers aren’t exact and it’s going to be 9-53%. Or maybe he’s way off and it’ll be 1% or 82%. The numbers aren’t the important part, it’s what’s coming that you can’t see but he can. And that’s ok. Not everyone gets to see the future, not everyone sees the same part of the future, and some don’t see any future.

Robotaxis are coming. Bet according to what you believe and quit arguing over the stupid stuff. You all sound like you’re standing on the side of the road arguing over which fork in the road to take when both end up at the same location. Big picture.

 

[Knightshade]

So @Green did tie HW4 (which is the version of the FSD Computer) to the configuration of the infotainment system in that particular implementation in the Model Y in his post.

I do think that's confusing, however. While there are undoubtedly some dependencies (the infotainment system needs at least XX RAM and YY GPU in order to display FSD visualizations, for example), I don't believe there's any specific reason the FSD system would dictate infotainment capability above those minimums.

Thus, I don't think "HW4 infotainment" is really a thing. It's quite possible another implementation (the CT for instance) may have HW4 and an infotainment system with greater specs.

I'm not sure how it's confusing though given as you say he specifically calls this out as applying to the Y? We already knew there's higher-end infotainment specs with the S/X than even the HW3 Ys had let alone the more downspec HW4 Y (which again makes sense since that's the only models with discrete GPU and that runs steam).

We don't know which of those two versions (Y or S/X) will be used in CT of course... have recent versions of it shown the rear screen we saw in original versions? If so that makes the S/X version more likely... and we don't know which will go in the refresh 3 either (though it'd be pretty weird if it wasn't the Y version in that case at least). They could even do some 3rd version, but Tesla tends to like fewer variants of parts, not more.

 

[Zaddy Daddy]

...and yet, the supercharger network does actually generate a profit.
How about waymo and cruise? whats the profit margin for them? I love how you use the term 'not a big profit generator' for something that actually exists, but hand-wave away the ludicrously loss-making cash-burning mess that is waymo and cruise because.... reasons.

• If waymo actually had the same number of driverless miles as Tesla I'd be worried (they are below 0.1% as I recall?)
• If waymo currently made a profit I'd be worried
• If waymo actually made their own cars and didn't have to pay a profit-margin to a 3rd party I'd be worried.

I am so not worried. I'm very bearish on FSD. I think robotaxi is many years away. I have zero doubt Tesla get their first. HD maps are out of date the nanosecond they are recorded, and LIDAR sucks in bad weather. There is NO answer to these criticisms of the ridiculous waymo model. They are doomed by design and will never get free of their local maxima.

There already are robotaxis. The issues Waymo and Cruise do have has nothing to do with Lidar and HD maps, I don't know why you keep bringing them up.

Local maxima? Cruise and Waymo have been showing improvements every year. They still have too many remote interventions, but that has nothing to do with Lidar or HD maps. Their software (like Teslas) should keep improving as they deploy to more cities (and they've stated they are observing this happen).

No, Waymo and Cruise's main issues will be cash-burning as you've mentioned. But just becuase they haven't reduced cash burn / vehicle yet doesn't mean they won't ever. Certainly it has to happen before investors lose confidence - which is why they are aggressively expanding right now.

 

[SO16]

Tesla has ZERO driverless miles.

It is premature for any company to be driverless.

 

[Tony73]

Tesla has ZERO driverless miles.

Technically that’s not true. Tesla has zero unsupervised miles -at least that were legally meant to be unsupervised.
Here’s the thing; Cruise and Waymo do a lot of remote supervision. They even have chase cars following on some of their rides.

 

[AndrewZ]

Agreed about the waymo model being doomed. Not so sure about FSD L-5 being years away (means 5-10 years). Just drove to NE on .4. Significant improvement. Only one takeover on a one lane both ways construction zone (out of the six times--3 zones times 2 ways). This was the first time that the car made it on the turn through Oklahoma City without a takeover. I'm even finding that I'm using it most of the time on surface streets around DFW--works great in almost all cases if you don't use the Nav system--urban only, trip always use Nav. Still urban surface, there is still a significant problem with map-vs-vision (weird lane changes, often turns one corner too soon, doesn't stay a fixed distance from the centre line, takes too many minor streets compared to more trafficked streets, wants to do the speed limit rather than a safe speed on those minor streets. Minor means residential with cars parked on both sides leaving only one lane for both directions).

Level 5 isn't 5-10 years away, it'll just be 1-2 years away for the next 5-10.

 

[Singuy]

Technically that’s not true. Tesla has zero unsupervised miles -at least that were legally meant to be unsupervised.
Here’s the thing; Cruise and Waymo do a lot of remote supervision. They even have chase cars following on some of their rides.

Seems like the bar is to not have a car hit anything after a tens of thousand of miles and not be intervention free to be a robotaxi.

 

[Buckminster]

First deluge firing coming up:

 

[cliff harris]

There already are robotaxis. The issues Waymo and Cruise do have has nothing to do with Lidar and HD maps, I don't know why you keep bringing them up.

Local maxima? Cruise and Waymo have been showing improvements every year. They still have too many remote interventions, but that has nothing to do with Lidar or HD maps. Their software (like Teslas) should keep improving as they deploy to more cities (and they've stated they are observing this happen).

No, Waymo and Cruise's main issues will be cash-burning as you've mentioned. But just becuase they haven't reduced cash burn / vehicle yet doesn't mean they won't ever. Certainly it has to happen before investors lose confidence - which is why they are aggressively expanding right now.

Does cruise or waymo have a solution for bad weather, when the LIDAR is absolutely useless? What is the point of HD maps when they are out of date immediately? Who the hell is going to install a robotaxi app for a ride hailing service that only works in decent weather? and only if there are no sudden roadworks or emergency diversions?
Waymo lose money on every journey. 'expanding' just means losing even more money. I hate to have to state the obvious but that is not a business model, its a pump-and-dump scheme that the executives responsible probably hope to quit and flee the company before anybody questions it.
check out the hardware:

Introducing the 5th-generation Waymo Driver: Informed by experience, designed for scale, engineered to tackle more environments

At Waymo, we’re building the World's Most Experienced Driver - our combination of hardware, software, and compute that powers vehicles to safely get people and things where they’re going.

waymo.com

3 lidar units and 3 radars, and cameras, all fitted to a high end jaguar i-pace for each car. Ridiculous expense. The chances of this being a commercial proposition that might be cheaper than uber is zero. Jaguar and the companies who make lidar probably love waymo.

Its LAUGHABLE to brag about waymo expanding to another ting ring-fenced area in another city. Big whoop. Tesla is collecting data from every Tesla on earth. Their dataset is unmatched, and at this point, cannot be matched.

If you want to pump the value of waymo stock, try twitter, or reddit, or anywhere not full of people who actually look into the details like we do here.

 

[SOULPEDL]

If any more FSD pessimism shows up, I may need to sell some shares and buy DITM Calls. (I know, I just became leverage free). But this noise is a tell tale sign of better Tesla coming very soon.

My instincts tell me that while Tesla was out window dressing FSD to license to other automotive companies, word got out within certain circles... NDA aside, the act of shopping Tesla FSD tech likely came with a few secret goodies coming soon. Scary stuff for many. And then I think somebody talked... and that was the plan.

Tesla knows the competition can't keep their mouth shut, but I don't mean publicly. Just the usual insider talk as real as the MMD and other coordinated Media attacks by Mr Market.

Yes, this is all made up in my head based on instincts, human behavior, and the last several pages of FSD FUD. This is an indicator, data. Trying to prevent a Call Options run ahead of V12 is a guess, but something's up.

 

[nativewolf]

Show me a profitable robotaxi company and then we can discuss FSD competitors. Elon has made this very simple and his prediction on robotaxi is spot on.

1. The car must be cheap and of abundance using the cheapest hardware.

2. The operating expense needs to be close to nothing, probably best to have the people carry most of the operating expense while Tesla makes money from the app by taking a percentage off the top.

3. The car must work everywhere, which means it will be able to take the most optimal route competitive to public transportation, self driving, or ride hailing. Every minute longer the occupant is in the car compared to regular driving = lost opportunity cost and kills demand.

Missing any of the key component above and Robotaxi is DOA as a service and DOA as a business. I know people are sick and tired to Elon promising FSD but honestly there are zero competitors out there with all or any of the above. This is why user experience is terrible, being stuck in a car 2-3x longer than it needs to be, why all of them are losing billions a year, and why Elon is 100% sure all will license, steal, or switch to Tesla FSD once it's solved as there's no other path to profitability. Wishing Elon start ride hailing today with geo fencing just to prove they also have some kind of L4 is foolish and I rather not see (-) cash flow on their earnings call.

I would say your first sentence contradicts itself based on any study of economic history, including an analysis of Tesla as you are insinuating that somehow Tesla has a profitable FSD solution...which they don't. Was Amazon a competitor in e-commerce? They didn't make money for a decade but they had a solution and were dominate. Uber in rideshare? Uber is a capital destroying monster but has a solution and maybe just maybe has found profits. We have no idea what Teslas solution will be, at this time it is years away from a product so purely academic. What we do know is that the industry is in the very infancy, two entities have products in the USA and several have products in China. Nobody knows how they'll get to market at scale but only these 2 have products. Tesla doesn't. Waymo is the clearest but I still wouldn't bet that we understand the strategy in full. When Tesla can deliver a FSD Robotaxi then we'll know how Tesla plans to attack the market and be able to discuss Tesla as a competitor to Waymo or Cruise.

Point 3 is obviously incorrect. That is a potential way for Tesla to differentiate but clearly absolutely not a requirement. Rideshare profits are in urban metro regions. Why offer anything else unless govt subsidies support it? Uber, Didi, Lyft, Grab has already profit tested the business models. I can't get an Uber to my home and I can't get one at my parents house. I see no reason for Tesla to offer a robotaxi in either location.

Without a doubt Tesla is the best positioned to actually build a robotaxi and software. Maybe unique in this regards. FSD can also become a very profitable product without robotaxis. EM has also said something else about robotaxis- he said it was critical for Tesla to have a solution early in the field.

My last post on any of these for a few years- right now its just been the same rehash as last year with the only update being Waymo and Cruise are moving to offer actual services at a faster and faster rate. For you investors this is just a data point to watch- how much of your investment thesis is FSD robotaxi revenue?

 

[Buckminster]

how much of your investment thesis is FSD robotaxi revenue?

33%
Bot=66%

 

[betstarship]

I'm surprised this hasn't been posted:

“We believe that anything that can bring fans to the site without having to drive their cars or making it more convenient for people is a real positive.”

https://www.sfgate.com/athletics/article/oakland-athletics-want-elon-musk-boring-loop-18272173.php

Pretty awesome to have a major sports team get behind Boring Loop!

 

[jerry33]

I'm not sure how it's confusing though given as you say he specifically calls this out as applying to the Y? We already knew there's higher-end infotainment specs with the S/X than even the HW3 Ys had let alone the more downspec HW4 Y (which again makes sense since that's the only models with discrete GPU and that runs steam).

We don't know which of those two versions (Y or S/X) will be used in CT of course... have recent versions of it shown the rear screen we saw in original versions? If so that makes the S/X version more likely... and we don't know which will go in the refresh 3 either (though it'd be pretty weird if it wasn't the Y version in that case at least). They could even do some 3rd version, but Tesla tends to like fewer variants of parts, not more.

I thought that in the original reveal, the CT had its own interface, different from either S/X and 3/Y. Now perhaps that's changed, but I haven't heard anything other than the original statement.

 

[SOULPEDL]

For you investors this is just a data point to watch- how much of your investment thesis is FSD robotaxi revenue?

More of it on Monday, but we’ll see.

 

[mongo]

I thought that in the original reveal, the CT had its own interface, different from either S/X and 3/Y. Now perhaps that's changed, but I haven't heard anything other than the original statement.

The screen and GUI is separate from the hardware driving it.
That said, CT likely has a completely different board set to support 48V.

 

[AndrewZ]

If any more FSD pessimism shows up, I may need to sell some shares and buy DITM Calls. (I know, I just became leverage free). But this noise is a tell tale sign of better Tesla coming very soon.

My instincts tell me that while Tesla was out window dressing FSD to license to other automotive companies, word got out within certain circles... NDA aside, the act of shopping Tesla FSD tech likely came with a few secret goodies coming soon. Scary stuff for many. And then I think somebody talked... and that was the plan.

Tesla knows the competition can't keep their mouth shut, but I don't mean publicly. Just the usual insider talk as real as the MMD and other coordinated Media attacks by Mr Market.

Yes, this is all made up in my head based on instincts, human behavior, and the last several pages of FSD FUD. This is an indicator, data. Trying to prevent a Call Options run ahead of V12 is a guess, but something's up.

What does FSD licensing mean to you exactly? To understand what this will mean to the financials, we need to understand what "license FSD" constitutes in the first place and so far we have zero details.

If "license FSD" means a suite of cameras and highway functionality, which Mobileye has already been doing for a long time and previously did with Tesla, I wouldn't expect much.

 

[SOULPEDL]

The screen and GUI is separate from the hardware driving it.
That said, CT likely has a completely different board set to support 48V.

At the local level, most chips are still 3.3V. So just a step down rail is needed to power whole boards (but I guess on board regulation is also possible, but that's a new part # then). The real 48V savings are with things like power seats, tailgate, windows and other small dc motors where the hot wires need to travel several feet.

 

[SOULPEDL]

What does FSD licensing mean to you exactly? To understand what this will mean to the financials, we need to understand what "license FSD" constitutes in the first place and so far we have zero details.

If "license FSD" means a suite of cameras and highway functionality, which Mobileye has already been doing for a long time and previously did with Tesla, I wouldn't expect much.

FSD License means others can't do it themselves in time before their entire platform becomes irrelevant. I'm not going to pretend to figure out the system or levels without more details. But to believe it's not happening is to dismiss Elon when he clearly states that one or more OEMs are interested. And it could even be at basic levels to eliminate 80% of safety risks, or could be their next platform and the whole computer with full sensor suite. But I'm dying to find out!