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

[BitJam]

I think mapping is fairly antithetical to what tesla is trying to accomplish with their vision based FSD. They are going for the ability for the car to fully handle roads / intersections it encounters which it has never seen specifically before. Maps in a sense are another source of "sensor fusion" issues. For instance there is a road a few blocks from me which in the past couple of months went from one way to two way, and another which was permanently reduced from two lanes to one. FSD has to be able to just deal with changes it hasn't seen before.

On the other hand, using what the car sees to update their map data for navigational purposes could be pretty useful.

Where I live some map or stored information is needed and not just from what the car sees. On my way home I turn off a 45 MPH divided highway onto a small 30 MPH road with an almost blind corner. The car thinks this part of the small road is 45 MPH which is nuts and very dangerous.

Despite the rapid advances in FSD beta, level 5 autonomy seems a long way off perhaps requiring new layers of processing to deal with small edge cases. Solving the last 1% may be harder than solving the first 99%.

 

[Singuy]

The main reason why Pierre Ferragu is not super bullish on robotaxis is that he thinks Tesla will need to work too hard to convince customers to book a robotaxi and will instead choose to take a driven Uber. I found this surprising as there a great deal of early adopters in the world that will jump at the chance imo. Do the non super bulls here agree? Most non super bulls tend to think that FSD won't be ready in the next 5 years.

Pierre had multiple points that gave him doubts and they are reasonable.

1. Will robotaxi demand be low in the near term? Waymo's program has demonstrated that demand for driverless taxi is low when it's too geofenced + not enough early adopters. This is currently true in Phoenix Arizona. Could just be that the tech is ahead of it's time so for the near term I agree with Pierre that if FSD was solved tomorrow, it'll take awhile before we have universal acceptance. When lives are on the line, people would like to wait a few years to see real world statistics and how often accidents happen. We understand that the media eats up any Tesla's mistakes and the first FSD accident will probably end up in the supreme court in the fight against Tesla's rollout given the spectacle it'll generate. I see major risk here.

2. Will robotaxi be profitable? Pierre mentioned google fiber being expensive. Ride hailing is also expensive. Uber being a software company makes only 38% gross margins on their revenue. This is absolutely abysmal when compared to other software makers. The reason is due to high cost on insurance and customer service, taking 40% of their total revenue. So it's going to be a long road ahead to reach profitability on ride hailing robotaxies. We are expecting revenue per customer to go down due to having no driver and to compete with people who own cars. However with price going down, we should see influx in demand which is where robotaxies should make it up. Taking marketshare from Uber is easy, however would half priced robotaxies make people want to ditch their cars completely? I think yes in the distant future, but not anytime soon.

So with Pierre I believe he thinks in the short term robotaxies will be a hard business to make profitable. FSD subscription/buy in he sees are much more profitable than to deal with the expense of the ride hailing network and the potential for a near term lack in demand. Also the price of Tesla's being 35k for Tesla to produce doesn't help with that profitability equation, but a simple 15k car with 2 seats or something may change things around.

 

[MC3OZ]

I think mapping is fairly antithetical to what tesla is trying to accomplish with their vision based FSD. They are going for the ability for the car to fully handle roads / intersections it encounters which it has never seen specifically before.

I don't see any conflict here, vision is the rule, maps are the recommendation.

When there is no map, vision is still the rule, then vision builds the map, to make it a bit easier for the next guy.

Say cars A, B,C, & D arrive at a stop sign, Car A sees and records the sign, Car B doesn't, all other cars do.
Over a large sample 80% of cars see a stop sign and 20% don't, Tesla can dig into where this 20% of false negatives is happening.

If 95-100% of cars always see a stop sign, the car can approach that area carefully assuming that a stop sign is probably there. In that case, even a false negative is safer, because the car is initially cautious.

if the stop sign has been taken away FSD will eventually work that out,

We have seen Tesla FSD presentations with cars driving around building up maps. If you don't think Tesla is using maps, then what were those images all about?

 

[MC3OZ]

Will robotaxi be profitable?

We have seen that Tesla regularly carries business units through the start up phase.

In most locations Tesla should be doing the insurance.

It is also likely that the new 25K vehicle with LFP and Model 3s with LFP are ideal for the majority of the fleet.

Using some private owner cars in the fleet reduces the need to built a large fleet fast.

Using off-lease vehicles to build the fleet lowers the fleet cost.

Tesla owns the Supercharger network, and electricity is a lot cheaper than gas.

The software development is still a substantial project but like all software, it is a single product deployed in multiple locations.

IMO the main risk to profitability would be building the fleet too fast, all Tesla needs to do is slowly build up the fleet in each location based on demand.
The fleet always needs to be larger than the peak demand, but that is an area where software can try to optimise the solution.

I still agree that there is plenty of use for FSD by a private owners, too old, young, sick, tired, to drive, let FSD do it. Pick the drives you want to do, let FSD handle the rest. Send the car to the vet to pick up the pet. Get the car to drive the kids to sport.

 

[stealthyc]

I don't see any conflict here, vision is the rule, maps are the recommendation.

When there is no map, vision is still the rule, then vision builds the map, to make it a bit easier for the next guy.

Say cars A, B,C, & D arrive at a stop sign, Car A sees and records the sign, Car B doesn't, all other cars do.
Over a large sample 80% of cars see a stop sign and 20% don't, Tesla can dig into where this 20% of false negatives is happening.

If 95-100% of cars always see a stop sign, the car can approach that area carefully assuming that a stop sign is probably there. In that case, even a false negative is safer, because the car is initially cautious.

if the stop sign has been taken away FSD will eventually work that out,

We have seen Tesla FSD presentations with cars driving around building up maps. If you don't think Tesla is using maps, then what were those images all about?

Training

 

[Ogre]

The main reason why Pierre Ferragu is not super bullish on robotaxis is that he thinks Tesla will need to work too hard to convince customers to book a robotaxi and will instead choose to take a driven Uber. I found this surprising as there a great deal of early adopters in the world that will jump at the chance imo. Do the non super bulls here agree? Most non super bulls tend to think that FSD won't be ready in the next 5 years.

Personally, I’m glad to hear a counter-point.

My feeling on Tesla is it is a great stock with tons of potential add-ons.

The base/ EV business is what it is currently valued on primarily and that business (Margins plus growth) is very likely to carry the day without any optional/ bonus revenue. It wouldn’t be a great stock with just that, but it would be decent.

Where Tesla becomes a fantastic investment is in the options/ not-priced in business lines. Pierre’s suggestion is that even with FSD being *just* a nice driver assist feature, Tesla is a still fantastic investment. He is a bull in-spite of his skepticism. As with many of the other “side businesses“, the value of FSD is not really part of the main thesis for Tesla.

It’s almost like owning the worlds fastest growing automaker and owning options to a giant pile of revolutionary start-ups.

* FSD/ Robotaxi
* Power Storage
* Virtual Power Plants
* Teslabot
* Tesla Semi
* Intelligent Insurance

Any one of these mini businesses under the Tesla umbrella might end up worth hundreds of billions of dollars. Some could be trillion dollar businesses themselves.

We don’t need them all to hit 100%. If FSD is merely a great driver assist package then we still have a giant bag of other fantastic start-ups to lean on.

 

[EVDRVN]

Read the agreement you signed when you applied for a permit to connect your solar to the grid. I'm betting it probably has language that prohibits the disconnection of your generation simply to send a message or conspiring to wreak havoc on the grid. If it doesn't, it's only because the utilities lawyers were not doing their job very well (protecting the utility).

That's not to say it couldn't be done anyway.

I can turn off my disconnect for any reason including working on the house. You would also be referring the the agreement they have already broken

 

[stealthyc]

Where I live some map or stored information is needed and not just from what the car sees. On my way home I turn off a 45 MPH divided highway onto a small 30 MPH road with an almost blind corner. The car thinks this part of the small road is 45 MPH which is nuts and very dangerous.

Despite the rapid advances in FSD beta, level 5 autonomy seems a long way off perhaps requiring new layers of processing to deal with small edge cases. Solving the last 1% may be harder than solving the first 99%.

How would I navigate it as a human the first time I do it? I would hopefully not need a map. I would drive it safely based on my view of the road and its conditions. In fact incorrect speed limit data from maps is a great example of the sort of problems maps offer currently.

 

[jerry33]

How would I navigate it as a human the first time I do it? I would hopefully not need a map. I would drive it safely based on my view of the road and its conditions. In fact incorrect speed limit data from maps is a great example of the sort of problems maps offer currently.

A human driver first looks at the map if he or she is not perfectly familiar with the destination to plan the route and/or look at current traffic conditions. After they have driven it a few times, then they would only use vision and would change the route on the fly as needed (construction, accidents, etc.), or plan an alternate route in the case of traffic issues. So maps will always be needed for route planning. Google/Apple maps or Waze are still maps and humans use them--just not in the same way that LIDAR based systems do where every inch has to be perfect. FSD should be using maps in the same way that humans do.

 

[Knightshade]

We have seen that Tesla regularly carries business units through the start up phase.

In most locations Tesla should be doing the insurance.

AFAIK Tesla doesn't actually underwrite insurance anywhere- they're basically a reseller/broker for other underwriters in a handful of states. In CA they're underwritten by State National Insurance... in Texas by Redpoint County Mutual... in Illinois by Midvale Indemnity... and in WA by Homesite.

Last I knew ARK has speculation about what a hypothetical insurance business could be worth if Tesla did actually begin being an insurance company instead of just a broker/reseller, but they've not actually done it yet.

As we've seen by how long it has taken just to jump the legal and regulatory hoops simply to be a broker/reseller, to the point that a couple years after announcing the business they're still only in a few states-- actually underwriting their own policies on RTs would.... take a while.

It's absolutely a smart eventual goal, but it's pretty far off.

 

[MP3Mike]

Some people are finally doing the math:

Montpellier foregoes hydrogen project for battery-electric buses, for now | electrive.com

www.electrive.com

However, these calculations now appear not to have panned out. Instead, Montpellier is foregoing the purchase of the hydrogen buses in favour of battery buses claiming operating costs for the H2 buses would be more than six times higher than for battery buses. Specifically, the city calculates it would cost them 95 cents per kilometre for the H2 buses versus 15 cents per kilometre for the battery buses.

But haven't totally given up:

However, they do not want to bury the hydrogen bus project completely: “We will do without hydrogen buses for the time being and will see in 2030 whether hydrogen is cheaper then,” added Delafosse.

 

[space_s3x]

All Tesla price targets on a single web page, created by Steven Bink:

Tesla Price Targets

Calculate the potential investment value of your Tesla shares, with price targets from the largest collection of TSLA analysts, institutional and retail! | When can you retire on Tesla stock?

teslapricetargets.com

 

[Duffer]

Regarding the proposed Solar changes here in California. I'm going to use the provided links and make my voice heard and I encourage that everyone else here does too (VPN's aren't just for watching sports ). However, with the unmitigated gall and sense of entitlement that PG&E has demonstrated previously (see below) I'm worried that the CPUC vote is merely a formality.

After being found liable for yet another major wildfire resulting in multiple deaths and many millions in damages, PG&E expected the CA legislature to grant them immunity. Apparently, the only thing politicians like more than money is staying in office and with the huge public outcry against PG&E the legislature denied the request for immunity.

Great thought!!, I'd LOVE IT if some superchargers in very high traffic areas had a way you could have your car cleaned- even detailed- while you charge. You could pay for a complete detailing - or any of a series of 10 - 15 minute service items that you'd get done over a series of charging sessions over time at stations equipped to offer that service, even if it is over a period of a year or two.

There's a company here in NorCal that does that at selected supercharger locations. They offer wash, interior/exterior detail, HEPA and cabin filter changes.

 

[ZeApelido]

All Tesla price targets on a single web page, created by Steven Bink:

Tesla Price Targets

Calculate the potential investment value of your Tesla shares, with price targets from the largest collection of TSLA analysts, institutional and retail! | When can you retire on Tesla stock?

teslapricetargets.com

Great way to find the clown show

 

[Big Time]

Listened to the Ferragu & Dave Lee podcast this afternoon. Ferragu seemed disinterested in even considering FSD in any sort of probable scenario with his forecasts, and he was firm on saying that using Autos alone, Tesla is a complete bulldozer that has absolutely no competition.

Him saying along the lines "If Tesla gets its expected percentage(Cant remember off hand) of the global market share, I cant even fathom who their competitors could be to fill the rest of the pie" was such a great quote. Autos alone, that's all we need, everything else is just icing on the cake with these margins.

 

[StealthP3D]

All Tesla price targets on a single web page, created by Steven Bink:

Tesla Price Targets

Calculate the potential investment value of your Tesla shares, with price targets from the largest collection of TSLA analysts, institutional and retail! | When can you retire on Tesla stock?

teslapricetargets.com

Does anyone understand the reason for the odd data entry for # of shares owned? It only allows you to set your share number in strange increments. Why not just let it be an integer? And what is the purpose of having a slider to adjust the number? It bothers me when I don't understand how something odd came to be.

 

[ZeApelido]

Does anyone understand the reason for the odd data entry for # of shares owned? It only allows you to set your share number in strange increments. Why not just let it be an integer? And what is the purpose of having a slider to adjust the number? It bothers me when I don't understand how something odd came to be.

Just PM me your share # and account info and I can take care of it.

Mod: I know you're joking, or at least you'd better be joking, but for the record don't do this ever again. --ggr, who is in the middle of reading Nicole Perlroth's "This is how they tell me the world ends", so I'm a bit sensitized to anything resembling phishing.

 

[Discoducky]

Does anyone understand the reason for the odd data entry for # of shares owned? It only allows you to set your share number in strange increments. Why not just let it be an integer? And what is the purpose of having a slider to adjust the number? It bothers me when I don't understand how something odd came to be.

Looks like it is being addressed...and it needs to go further to the right

 

[Tyler34]

All Tesla price targets on a single web page, created by Steven Bink:

Tesla Price Targets

Calculate the potential investment value of your Tesla shares, with price targets from the largest collection of TSLA analysts, institutional and retail! | When can you retire on Tesla stock?

teslapricetargets.com

Seems only @The Accountant made it on the list from TMC.

 

[EVDRVN]

Regarding the proposed Solar changes here in California. I'm going to use the provided links and make my voice heard and I encourage that everyone else here does too (VPN's aren't just for watching sports ). However, with the unmitigated gall and sense of entitlement that PG&E has demonstrated previously (see below) I'm worried that the CPUC vote is merely a formality.

After being found liable for yet another major wildfire resulting in multiple deaths and many millions in damages, PG&E expected the CA legislature to grant them immunity. Apparently, the only thing politicians like more than money is staying in office and with the huge public outcry against PG&E the legislature denied the request for immunity.


There's a company here in NorCal that does that at selected supercharger locations. They offer wash, interior/exterior detail, HEPA and cabin filter changes.

Thank you at to all for the support, I'm not sure if everyone realizes how significant this is and how detrimental. Smashing the largest solar market in the US will impact all US solar markets and really impact Tesla negatively. PGE is a criminal enterprise plain and simple and it needs to stop.