-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable
- Thread starter AudubonB
- Start date
SAPharmD
Member
Maybe everything will cool down a little by Monday (hopefully)...
Ashok said, "This is the beginning of the end." I don't know about you, but it definitely perked my ears up.
And you have to admit, Musk's bullishness is different this time. Elon appears to be betting the whole company on full autonomy. Spare no expense.
He hasn't done anything like that before. Before, he would say, "I think it's going to happen this year." Now he is saying, "Spend billions to make this happen tomorrow."
Yes, Tesla has more data on battery degradation than anyone on Earth, and yet we have absurdly short battery warranties that kills BOTH used Tesla sales, and resale value, over the long term.
Why?!?
Tesla should step up and increase the HV battery (and powertrain) warranties across the board to a full decade and 200,000 miles, NOW.
This is a ticking time bomb for Tesla, and it gets worse with each passing day.
Social media is replete with five-figure repairs estimates to replace batteries on older Model S's--which is about or more than the value of the car. That sends a terrible message to those that buy used cars, which in raw numbers are more than those that buy new cars IIRC?
For such a smart company, I fear that Tesla often makes some pretty stupid mistakes, and this greatly harms Tesla's image, and our investment.
I'm not saying Tesla will definitely want to own the whole fleet. But I believe they could pull it off if that's what they decide.
If robotaxi works, Tesla can raise the cash to do whatever it takes. Tesla could do another public offering at $3000 per share.
Also, Tesla can't snap its fingers and make 10 million cars appear. It would take a few years to ramp up. In that time, Tesla would be generating huge revenue on the robotaxi vehicles it can produce. The fleet would be a cash machine that keeps building and building each year, helping to fund an even larger fleet.
It could be done.
Krugerrand
Meow
Well, I just tried my left turn for the third time on 12.4 and it failed. But 2 of 3 is still better than the 100% failure rate on 12.3.
The frustrating thing is that this is so much easier than Chuck's unprotected left. All it has to do is get into the turn lane. But the car is scared of this particular lane. It's like the car thinks that lane is haunted.
Artful Dodger
"Neko no me"
Yeah, I don't think so. Have you run the numbers? Even at $3,000 per share and $18,000 per robo taxi, building 10 million robotaxis per year would result in dilution of about 60 million shares per year. Are you sure Tesla wants to do that? Further, are you sure that shareholders would want that?
Artful Dodger
"Neko no me"
Do you mean 12.3.4? Please be accurate, this will save a lot of confusion in the future when looking back on this archive. TIA.
We need to wait and see what is revealed on 8/8, including any details of business plans that might be revealed.
it is clear to me that adding a wheel with driver feedback adds some cost of a Robotaxi, but there may be compensating additional costs. So a Robotaxi is a specialised vehicle to some extent.
it isn't clear how much of the Robotaxi fleet Tesla intends to own, private investors or corporations might purchase a fleet of Robotaxis to operate in the Tesla network. That is one model, as is Tesla owning the Robotaxi. Additional models are private ownership of Model S/X3/Y and Tesla also owning some of those vehicles.
Initially Tesla might prefer to sell Robotaxis to private owners rather than own them outright, but it will probably be a mix.
I expect the initial line at Austin to produce around 5,000 Robotaxis per week, I also think that a Robotaxi might be the first Gen3 vehicle produced in Mexico and Berlin. I can see Tesla might be at a run rate of around 1 Million Robotaxis per year in 3-5 years time. But at that time I also expect them to be producing a number of additional Gen3 vehicles, a lot more Megapacks and be earning some income from the fleet of Robotaxis.
Once they get to a certain fleet size, the Robotaxi fleet itself pays for fleet expansion.
it has always been clear to me that the Gen3 product ramps will be steep, the next growth wave will be a big wave. Elon typically thinks as big as he can afford to think.
Last edited:
need my CT to get to Zero gasoline ... even got a battery powered outboard (torqeedo)
for my sailboat .. once CT is delivered no more direct gasoline purchases for me my 10 year goal is almost complete three years early
GFY exxon mobil... come on CT lets go
for my sailboat .. once CT is delivered no more direct gasoline purchases for me my 10 year goal is almost complete three years early
GFY exxon mobil... come on CT lets go
Yep, the tone from Tesla has changed, they are moving forward with FSD and RT's physically now. Revealing the RT in four months, dropping sub prices on FSD, moving FSD out of beta and releasing it to everyone, dropping prices and margins on cars to sell as many as possible in order to widen the market for FSD customers.
Sure Elon has been "crying wolf" on FSD for a long time now, but Ashok's tweet should not be taken lightly. The Robotaxi's are coming, like for real this time.
No, I haven't run the numbers. I pulled $3,000 out of thin air. If $3,000 per share isn't enough, they might go for $5,000. Or $10,000.
You don't really have to run the numbers. You just need to understand the value of each robotaxi Tesla produces. I've seen estimates all over the place, but even the very conservative numbers are huge.
My point is that a robotaxi produced by Tesla for $18,000 will be worth many, many times that amount over several years. So investors will be willing to fund the manufacture of those vehicles, whatever the cost, as long as the promise of future profit is there.
Again, I'm not saying that's the route Tesla would take. Tapping the capital markets is only one way to fund robotaxi expansion. I'm just saying it's one of the possibilities.
Last edited:
Yes. 12.3.4. Sorry.
I had not seen this posted here. KBB estimates of Q1 US EV sales by brand and model.
Interesting that the Cybertruck by itself already outsold Lucid and it should surpass the Rivian R1T in Q2.
Interesting that the Cybertruck by itself already outsold Lucid and it should surpass the Rivian R1T in Q2.
2daMoon
Mostly Harmless
Tesla doesn't need the fleet, though I expect them to have one, as well as many other businesses providing Robotaxis for hire.
Tesla's ride-hailing App and related services will be what coordinates the actions and transactions of every taxi tied into that service.
These taxis will only be Tesla vehicles. Either built by Tesla, or built by others with Tesla licensed technology.
Whoever is managing such a business pairing people with on-call transport, and processing the payments will become stupid profitable.
Owning and operating the Robotaxi hailing business is much more important as a profit center than Tesla owning Robotaxi vehicles.
TAAS is where Tesla will rule, even if they were to manufacture and give away the Robotaxis that are knitted into that network, they will have crazy margins.
And no, they won't be giving them away.
Last edited:
ChatGPT and other AI systems make use of reinforcement learning with human feedback (RLHF) where people can indicate when the system did something wrong to fine-tune future behaviors to avoid the mistakes. Video data collected from Tesla fleet actively using FSD is easily annotated for when Autopilot was engaged and when the driver disengaged and controlling.
With additional labeling (both human and automated), Tesla can then identify any portions of disengagement clips that 1) FSD should avoid doing in the future and 2) how a human fixed the mistake. The former (1) reduces the current neural network's desire to do some action, e.g., overtake crossing double-yellow, effectively increasing the likelihood of other actions, e.g., stay in the current lane or slow down. The latter (2) even if is not ideal driving behavior of correcting a mistake, e.g., swerving back into the original lane, is still useful in not only learning how to correct but more importantly "why" such as boosting the internal neural network signal of "is this an appropriate time to overtake?"
RLHF might not be "pure" reinforcement learning because it requires humans in the loop instead of some fully automated system, and it can be expensive or limited by how many humans are providing the necessary feedback. Tesla deploying a free trial and lowering the subscription price could effectively be Tesla's way of paying for this very valuable training data, and delivering the latest FSD version to the fleet has other costs but is worthwhile to identify the mistakes that the newest neural networks are still having problem with to then collect and correct.
Tesla could finance a substantial chunk of the fleet given the relatively straightforward business of TaaS and, a guess here, there may be tax strategies that add offer some advantages.
Musskiah
DisGruntled
Tesla now putting up a South American offensive in Chile while also penetrating new markets like India.
www.google.com
Meanwhile, Elon is putting up his fight in Brazil; the dude sure does know how to compete on multiple levels simultaneously. I give him a lot of credit. He gives them a ton of fodder. Never a dull moment in this saga...
Tesla's first Model 3 and Y land in Chile as South American offensive begins
Tesla's first Model 3 and Model Y units have landed in Chile as the company's efforts to begin ramping up EV sales in South America have officially begun.
www.google.com
Meanwhile, Elon is putting up his fight in Brazil; the dude sure does know how to compete on multiple levels simultaneously. I give him a lot of credit. He gives them a ton of fodder. Never a dull moment in this saga...
Similar threads
- Locked
- Locked
- Locked
