I hope this video helped you get brighter about Tesla! My website: www.HerbertOng.com See the MOST comprehensive resource for the $TSLA Investor Join this channel or my Patreon to get free access to 15+ modules of TSLA investor resources Join this channel: th-cam.com/channels/4DBLlq1x0AKmip1QJUcbXg.htmljoin Join my Patreon: www.patreon.com/brighterwithherbert Thank you for supporting this channel! I really do appreciate your likes, subscribes and comments. Let's get brighter! Herbert
Don't forget, you have to understand the take rate of ride hailers. How many miles can be collected by the 100k miles? If you have a vehicle in a small town where there is only 200,000 miles per year of hailing, you will be hard pressed to multiply out anything more than 3 vehicles for that town. Large cities, yes, there is better market for it. But you can't calculate beyond demand. However, you will need more than 3 vehicles because of concurrent fairs. You'll be hard pressed to get that 100k out of each of those 3 cars. Consider too, you can place the cars where you want them, but you can't choose the destinations. If they go out of your market, some miles will be lost. Fortunately, once you have the data of the actual demand, you can then find the actual take rate and even optimize location availability based on that data. Tesla can do this. But to predict it is much difficult.
Another real question is, even if disengagement is extremely rare, what does a robotaxi do if there is no one in it? You can't have even one disengagement and have no solution for that. Otherwise, it may end up in bad situations. I sense there may be some degree of remote operation if they come out before they have zero disengagements in all situations.
I got a FSD demo two weeks ago. I sat in the passenger seat while a Tesla rep programmed a test drive rout of a few miles. The car would have done the whole drive itself except the rep showed me how easy it is to disengage and reengage FSD. It was amazing.
You're lying. It needs someone in the driver seat. And it's not amazing. It can kinda do the simple stuff that a 10 year old human can do. It cannot do anything more difficult.
This seems like the worst period. Even the market are now very unpredictable. Started investing recently when the market prices were a bit high,today I am more than 60% down!
Don’t be confuse buying the dip in a bear market, with guaranteed future returns. Just because that company is down 60%+ from ATH does NOT make it a sound long-term investment. Make sure you’re investing in great companies. kudos to Devion Jane
Her technical analysis is excellent and her interpretation/projections of the market are so accurate I sometimes ask myself if she is a human The point she is the perfect trader to follow for advice and guidance
I think Teslas robotaxis will come faster than most think. The progress is not linial anymore but exponential. More and more compute, more cars and better build software . I think we will see the first this year.
Most commentators have no clue as to how advanced Tesla is on self driving. They have never been in a Tesla and certainly never experienced the most recent versions. I discount their comments as being from a ‘no knowledge’ perspective.
Re: FSD Perceived as "Good Enough". I've also used it whenever possible, since Mar'22. But the real issue is how much Jonesing do you have when it suspends you for 5 days? I've learned to be MUCH move avoidant of letting that happen, especially since End2End Neural Networked V12 installed! Got V12.3.6 yesterday & can't wait to try autopark/summon! I'm thinking that V12.4.* will be another level up.
In Toronto Canada and Vancouver Canada, for example, it still has work to do, when comparing other states in the USA. I try not to get my hopes up but I presume it'll be truly dialed in by mid to late 2025. IMHO
As FSD 12 gets closer to 13 (a teenager now) it is at a point where as an interim fix such as Cruise's off site human monitor. With a single intervention in several drives a single monitor could handle multiple vehicles, receive a notification when a questionable situation arose, and handle the crisis. With one intervention per hour and remedying the situation in 3 minutes or less a single monitor could handle 20 cars at a minimum, so allocate 1/15th of a salary for each car, and that worker is easily covered. That covered number of cars would grow with each advancement of FSD until there was no longer a need for monitoring.
As a software/hardware developer of robotic type systems - multiple sensors (lidar radar vision) can create more problems than it solves. Multiple signals need hugely more compute, to solve via AI. Humans use (mostly) vision to drive...imagine if we had a radar screen, and a lidar screen to look at, as well as our vision, when we drive.`First principles thinking says that isn't a good system. Too many inputs can be worse than a top level, single input, (vision) , which is how we blood-bags drive....adding multiple cameras solves the "additional sensor" issue, in my humble opinion....
My take is that multiple sensors types are not inherently bad and the training and inference compute is not that much more - but it is the quality of the sensors is of massive importance. With radar and lidar Tesla was getting massive amounts of false positives. The lidar and rader would see things that are not there asking the driver to take over. Cameras don't really have that issue.
Re: Vision-Only FSD "Sufficient"? I've been following ADS, since the old DARPA Challenge days, from which I concluded that LiDAR lacked sufficient resolution to complete with vision. What's your engineering take on that? I also agree with Munro, that FLIR might be a significant help in highly inclement/dusty conditions. Would FLIR's cost be justified to accommodate those highly rare scenarios or would vision just reducing speed, when temporarily obscured, suffice? Meanwhile, Re: FSD Perceived as "Good Enough". I've also used it whenever possible, since Mar'22. But the real issue is how much Jonesing do you have when it suspends you for 5 days? I've learned to be MUCH move avoidant of letting that happen, especially since End2End Neural Networked V12 installed from the old heuristic coded V11! Got V12.3.6 yesterday & can't wait to try autopark/summon! I'm thinking that V12.4.* will be another level up, as it seems that NN "adaptations" may actually have started to deploy, mid-revision. I've "detected" such corrections running past the same "troublesome" points in my drives. Is that what you're seeing in V12, vs V11?
This lines up with the “combinatorial explosion” problem Jon Vervake has discussed in AI research. The fewer sensor modalities means greater focus concentrated on one.
Also if it is too foggy or raining to hard or blizzard it white our conditions or in a hurricane or tornadoes it may be best not to drive the autonomous BEV (Tesla RobiTaxi or CyberCab). I was thinking using otter sensors when needed might help but you probably don’t want to be around humans driving in those extreme weather conditions. Thoughts?
Well you humble opinion is mostly correct. The problem can become a "three body problem", impossible to solve, in fact we use filters to avoid some of the problems.
People have become so desensitized to lying and deception because we are surrounded by lies, liars, and misinformation and it creeps in at many little ways.
At 9k hours in a year and at $1.00/mile, 50 miles/hour, utilization rate 70%=$315,000.00 revenues minus cost/year=AMAZING! At just 40% utilization: $180k revenues minus costs.
I live deep in the woods of western Massachusetts, our roads are terrible. FSD does great on our no line roads. It does not do well with our potholes, you can easily total your car if you’re not paying attention.
There has to be an additional $4.00 surcharge per trip on top of $1:00 per mile Tasha’s estimate to cover short trips. Assuming 15 trip per day = $60 per day for surcharges alone. This amount should cover tesla’s monthly payments including the FSD subscriptions , insurance , maintenance and electric charges leaving charges for miles driven as pure profit.
Pre-covid inflation Elon stated that Tesla could run a complete (all costs) robotaxi service for $0.18 per mile. Obviously things like waiting times and repositioning while empty would have been factored in. Bump that up to $0.23 per mile to allow for the approximately 25% inflation bite we've suffered. Short term Tesla will be able to charge $2 or so per mile. Just make it cheaper than Uber/Lyft plus average tip. Once there are enough robos to replace driven taxis then the rate is likely to start coming down. The magic charge per mile is about $0.60 per mile which makes it cost about the same as driving one's own car. Tesla, in order to fulfill their mission statement, needs to price private ICEV ownership out of the market. Selling for $0.60 with a cost of $0.23 is a 160% profit.
100k a year is 274 miles a day. Depending on charging infrastructure (auto charging) that seems a bit high. For longer trips it likely a car can hit that, but a place like SF being 7 miles by 7 miles, even when the car is carrying someone, it will mostly be sitting in slow traffic or stop lights.
Tesla has communicated that the gross profit per robotaxi would be 30 000 usd per year. So that's the number I use but I guess it could be higher at least from the start.
That’s why he did it…. Why do u think Tesla factory is actually owned by Tesla US… Because in return for his nod of approval. He hands over the keys to how things work and how he lays it out Elon technically came in and gave China an actual right to say they are a competing auto industry. Elon did that…. Single handily…
There is so much focus on August 8th for FSD which I am excited for but there will be an earnings call in July before that how will they handle that without stealing the Thunder from August 8th.
v11.3 is the UI/UX version. there is a different versioning for FSD though it can be confusing at times! You can check what FSD version you have in the software section of settings in your car.
She dodged the questions on how the vehicles are charged. 100k miles per year is 273 miles per day or 1.5 charges. Who, how, and where are you going to charge them? With Tesla slowing down supercharger deployment especially.
The range of a model Y is 260. I can see them adjusting the batteries to last one full day. They will charge overnight at a Tesla service station where someone will also clean them. That's not hard.
They'll be charged by remote, by same group of people who remotely monitor the auto-bots and get them out jambs that they will occasionally find themselves in.
They're going to need a depot to park and maintain them at in each major territory. I would imagine the chargers would be there. That's one charge at night, one partial charge in the day at about 20-25 minutes.
The margins on the robotaxi will be high, but the cost of the fast growing datacenters needed for FSD will also be high. I wonder how these costs are applied in the very positive financial modelspoor.
Regulatory problems are not related to self-driving, rather to local city/county laws related to cab services. You see that happening now in California where RoboTaxi's has to comply with local laws. I'm not saying that can't be solved, maybe using Grok to know local laws, but it's not just as simple as no regulation. There actually will be a myriade of local regulation.
There will always be regions of strong incumbent resistance to new, better ways to do business. Tale as old as time. USUALLY the better economic model triumphs. The bigger the disparity, the better the and faster the conquest...and the disparity here appears pretty huge.
@@737smartin 100% agree, but it does not change the fact that RoboTaxis will have to comply with local zoning laws, in the same way as human taxis. This will require local knowledge to operate, which could come from Grok.
I think Monte Carlo is most useful when you have some real data from a small data set, or a small amount of data from a large data set, but not enough to make statistically significant conclusions. It allows you to extrapolate when you don't yet have statistical significance. If you have no real data at all, and are just making everything up based on guesses, Monte Carlo just gives you more guesses. It does nothing to improve the quality of your guesses. Maybe it allows you to explore how sensitive outcomes are to changes in certain variables, but that's it. I don't think it makes Ark forecasts any more dependable.
People are only looking at the supply side to calculate the possible revenue/year. they forget that if all Tesla’s are offering taxi services their will not be enough demand for that many Tesla’s.
Regulators do not intentionally try to prevent companies from releasing products and services, or else there will be no use for regulators. People think erroneously that Tesla’s path to regulatory approval is like going through the eye of a needle.
I see fsd being solved in the next 12 months, but it will be a few years before robotaxi takes off. Jmo. The average person is pretty skeptical about autonomous driving, they will have to see it over and over before they are convinced it’s safe.
I was in SF a couple weeks back. There were a lot of Waymo robos driving around. I doubt there will be a shortage of willing passengers starting week one. The skeptics can come around later. There will not be enough robos to service everyone at first.
To be approved as a taxi driver FSD does not need to be far better than any human driver, but just equal to any licensed taxi driver. The only caveat is that they have insurance equal to or better than taxi vehicles.
I believe BYD will partner with Tesla and that's the reason FSD got approved in China. Tesla needs a strong EV partner to take the entire marked rapidly
As a separate safety issue, FSD cars will probably require at least a mannequin be in the usual driver seat. Other drivers and pedestrians might become distracted when seeing no driver and very likely cause accidents.
@@didier_777 Yeah, I believe some people are crazy enough to jump in front of a slow moving FSD car just to try and collect insurance. Same thing with other drivers who might quick stop in front of a FSD car just to try and collect insurance. In a crazy world people do crazy things.
@@Mrewink5 @Mrewink5 The difference is that if anyone does that there are 8 cameras that will record everything so I don't think people will be dumb enough to try to scam the insurance companies with video footage. If anything those people will stay away from Teslas.
Monte Carlo method is great, averages out range of guesses of different inputs. Intended for realistic confidence interval. Need sensitivity analysis too to determine influential factors in model. Beware of GIGO garbage in, garbage out.
To increase TAM, need to aim for much lower rates than Uber. Tesla TaaS won’t merely replace Uber / taxis. That’s tiny market. It’s got to replace vehicle ownership, even substitute for public transport. Big challenge is to get to shared travel - almost always we travel like a herd, same direction. That won’t be on re-purposed private vehicle. Next generation can’t come soon enough, dedicated taxi design, for future, not past.
Obviousy you don't live in a city. Robotaxi can't substitute for public traffic. If everyone that took the subway in New York City took a Robotaxi instead to work I don't think that many would make it. That would be one increadible traffic jam.
@@arseliopereira2541 trains only work well as you describe, are poor service everywhere else. When younger, I found a pushbike to be quicker. Bus is poor everywhere in my experience. Trams are poor compromise. But it’ll be a challenge to get TaaS to work as I propose. Who’s up for it?
@@iandavies4853 I'm not suggesting that it's better to ride a bus than in a Robotaxi. I'm just saying that there isn't enough road capacity for eveyone to drive into the downtown of most cities in their own vehicle.
@@arseliopereira2541 exactly. Hence need for shared vehicle, not single occupancy. I see lots of huge 50 seat buses with 2 passengers (off-peak). Still same convoluted route, picking up no-one. CBDs are different problem. We all talk to our own world. That’s not mine.
Googled fsd for Canada...at least one province has outrightly banned it."Earlier this month, changes to B.C.'s Motor Vehicle Act quietly came into effect, prohibiting the use of fully automated self-driving vehicles.Apr 22, 2024"
Robotaxi prices won't be super cheap simply because costs are less. They will charge whatever the demand allows, just as they did with the Foundation Series gimmick. All they have to do is undercut Uber, they don't have to cut the price in half. In fact they can charge the same price because people, especially women, will choose the Robotaxis to avoid dealing with an unknown human being.
True value (productivity gain) requires change in behaviour, so RoboTaxi will have much more impact in low income countries (China, India), where people does not own a car, and with RoboTaxis can afford to drive. In high income countries (USA, Europe) where people already have a car.
As much as I like Herbert and I really do, I cringe just about every time he shares time with Larry because he keeps cutting in, doesn’t let Larry finish his sentence and/or thoughts, never mind expanding on his (Larry’s) thoughts. Larry at 78, has a lot of experience but still a young and furtive mind albeit with the restraint of innumerable years in business experience. This guy, Larry, is just such a resource of knowledge and measured thoughts of future outcomes.
ROBO-TAXIS ARE INEVITABLE - Just so much better than human drivers. FSD 12.3.4 is amazing, not 100% all the time but very close. 8/8 will be interesting!!
I'm glad I'm not the only one who was SHOCKED by not only China's acceptance of FSD, but also the impact it had on putting TSLA back on the Wall Street track. Thanks Larry and Herbert!
When it came to Tesla as a private robotaxi: did Larry Goldberg only think once about the fact that it requires interior and exterior maintenance after every single ride? At least the essentials? Just like every taxi driver can only survive if his vehicle is tip-top and in very good condition? Otherwise no one would get into someone else's car that wasn't absolutely clean.
Uber drivers do between 140 and 200 per day. At 100,000 miles per year, the Robotaxi is driving 274 miles per day. It doesn't account for any downtime for repairs or other maintenance that need to be done for a car driving that much. The car needs to go to the nearest Service center and good luck waiting for parts to be there. Overly optimistic models by Tasha without explaining any detail. Better analysis is done by Cern Basher and has a ton of realistic assumptions.
Chances are Musk is overestimating the pace that FSD will be adopted, assuming it reaches level 4/5. $100k is the same price that NY charged for taxi medallions. At the peak there were 300,000 taxis in the US. Very few people will dish out $100k. Tesla needs a sub $25k car for the masses.
The robotaxi/cybercab doesn't even need to be anything different than what they are already producing. Just being able to summon a regular model 3 or Y would be better than UBER. I don't need the driver. I can drive myself. What I need is for a car to show up at my door when I want it. Using "summon" through the Tesla app I could have any one of the 6 million Teslas currently on the road if they are put into the "taxi fleet". The future "cybercabs" in the densly populated cities don't need a dash or steering wheel or any of the instrument cluster. Think about that.... You could instead have a large compartment with rear facing seats where the dash/steering wheel is currently located and the front facing rear seat in the back...amazing leg room, seats 6 VERY comfortably in a car the size of a VW Golf.
Hm, in the US maybe, in Europe I doubt ppl will not share. Plus it is likely that TSLA will provide a rating system like Airb&b. Apart from that, would you not buy a car that would pay for its self, even is it takes 2, 3 or 5 years..?
@@heythave yes and what..are not the average BEV a punter of the future ? Remember the problem with the market now isn’t just interest rates it’s that we’re no longer in the market movers stage but the “laggards stage” with BEV’s ! How many of your acquaintances say they hate Elon and won’t buy a car from him ?
They will be making cars to sell and others to keep in their fleet. At first maybe it's one of 20 cars that will be robotaxi and at any point they can just ramp up as demand increases.
An issue (maybe temporary) could be pushback by Taxi drivers/Uber drivers. It's a lot of jobs which could disappear, a very old industry. Consider how much taxi drivers battled against Uber. Now consider Taxi's AND Uber against robo-taxi?
People wonder how Tesla could be worth a significant portion of worlds GDP. That's what happens when you convert unpaid labor (every unpaid driver) and convert that into revenue
Why is nobody addressing the demand side ? US have like 1.5 million uber cars. And even if tesla take like 1 million cars, even then it would not make much profits.
Well one big factor in the demand side is the cost. Electric miles are cheaper. Autonomous miles are cheaper. Combined they will cut Uber out of the market. But the big change will be people opting to use Transport As A Service rather than buying a car.
Uber is expensive af. If robotaxi could take me to work at the same cost I pay in gas but still less than an Uber I’d pay it so I can chill and read while it drives me
@@codingispower1816 i agree. But for that the cost of robotaxi has to be much below 1 dollar per mile. Also it might take people a lot of time to adjust to driverless cars. Also there might behuge resistance from current uber drivers and government because of huge loss of employment.
Autonomous taxi companies and investors are purposely pushing for higher taxi driver wages to squeeze human labor competitors out of the region by making operation costs greater then profitability. Once they force a minimum wage increase (under the guise of higher labor standards for drivers) they force a region out of the market. This shows then to have on hand a solution, their autonomous products. And for nearly the same cost as a human driver these robotaxis can be deployed in a region saving clients money, reducing co2, relieving congested roads, safety of ai over human error. Sketchy conspiracy and easy way to force ai over human by increasing labor cost at the government level.
Skeptic conspiracy theory: Talking to uber lyft fulltime drivers, they hadn't really tried to raise minimum wage, many were fine as they were. Now in the twin cities MN there's a push to raise taxi driver wages, and this has caused these services to plan on leaving the region. So if it's not the taxi drivers urging wage increase who is? Possibly the proponents push for higher taxi wages, are the investors of these AI robotaxi services? Who's to gain the most from current taxi services to withdraw from a region when costs are greater then profit? A competitor that wants to lock in a profit margin while squeezing out the human labor competition. Because no one invested in a company that only breaks even. Ethics and morals and client savings are just advertisement for greedy investors. It does seem very sketchy that at the same time taxi wages are being debated, an autonomous competitor would be ready to deploy with in the same time frame as services withdraw from a region. I am concerned there's some dark dealings our politicians are dabbling in for personal profits.
Dear Tesla, keep ramping up Cybertruck, Go pedal to the metal on new lines for CyberCab, Keep the first 1M you produce and start up the Fleet, Never stop building and take care of your employees. Your the one baking the pie, only normal that you get to eat it too.
Tesla’s in good condition, but do you really think they could afford to pay for 1m vehicles in one go? I’m sure they’ll be selling in excess of 50% of the cars off the line. Bear in mind the risk of building monopolies - Tesla won’t want to be faced with action on that. The pie is vast - one company doesn’t need to have it all.
What's the truth?!?! Tesla ALSO fired the destination charger team. Hotels and condos IN THE MIDDLE of installs are on hold, the work stopped, and emails to Tesla are bouncing!!!! Is Musk ok?
Ive interviewed Tasha and many other institutional analysts already. These videos are in the library. I interview new guests regularly including CEOs on a weekly basis.
And what prevents government and unions from featherbedding and passing laws which require driver in the who may or may not do anything but add cost? Musk and the public are not facing the fact that Musk is smaller than the government. His only option will be do what has been done for ever all over the world and that is to pay off politicians. I suspect he would walk away before dealing with that.
Ok. First the video talks about Tesla takes most. Then the video talks about why China invited Tesla and their FSD into China. I can see the conflict between these two views, can you?
If you can get $100k per year per car, why bother selling the cars at all? Just keep them all, shutdown the car sales business, and become a taxi company.
Just a philosophical question. Why should the citizens of America treat Elon's robotaxis with respect when he shows such disrespect for so many of America's citizens? (including employees)
We shouldn't show them respect and we will fuck with them to no end. Not to mention the number of highschoolers who will endlessly troll autonomous anything. It is gonna be a TH-cam clip gold mine.
Hertz also operates as a pseudo car dealer. Buying from the OEM,s at large discounts. After a short while they sell them for maybe a profit or a small loss, large part of their business model. With Tesla dropping prices to make them more affordable for the masses, Hertz took the decision to sell as the value for their business model was dropping
I hope this video helped you get brighter about Tesla!
My website: www.HerbertOng.com
See the MOST comprehensive resource for the $TSLA Investor
Join this channel or my Patreon to get free access to 15+ modules of TSLA investor resources
Join this channel: th-cam.com/channels/4DBLlq1x0AKmip1QJUcbXg.htmljoin
Join my Patreon: www.patreon.com/brighterwithherbert
Thank you for supporting this channel! I really do appreciate your likes, subscribes and comments.
Let's get brighter!
Herbert
Don't forget, you have to understand the take rate of ride hailers. How many miles can be collected by the 100k miles? If you have a vehicle in a small town where there is only 200,000 miles per year of hailing, you will be hard pressed to multiply out anything more than 3 vehicles for that town. Large cities, yes, there is better market for it. But you can't calculate beyond demand. However, you will need more than 3 vehicles because of concurrent fairs. You'll be hard pressed to get that 100k out of each of those 3 cars. Consider too, you can place the cars where you want them, but you can't choose the destinations. If they go out of your market, some miles will be lost. Fortunately, once you have the data of the actual demand, you can then find the actual take rate and even optimize location availability based on that data. Tesla can do this. But to predict it is much difficult.
Another real question is, even if disengagement is extremely rare, what does a robotaxi do if there is no one in it? You can't have even one disengagement and have no solution for that. Otherwise, it may end up in bad situations. I sense there may be some degree of remote operation if they come out before they have zero disengagements in all situations.
I got a FSD demo two weeks ago. I sat in the passenger seat while a Tesla rep programmed a test drive rout of a few miles. The car would have done the whole drive itself except the rep showed me how easy it is to disengage and reengage FSD. It was amazing.
Absolutely. And it’s just the beginning
You're lying. It needs someone in the driver seat. And it's not amazing. It can kinda do the simple stuff that a 10 year old human can do. It cannot do anything more difficult.
@@dynomotivedyno9194 lol.. and that comment is what ignorance looks like
Nice. Cool to hear this
@@dynomotivedyno9194 Total moron.
Thanks Larry, you are so much on the mark in your assessment of Tesla issues, including their FSD program! Appreciate you so much!
This seems like the worst period.
Even the market are now very unpredictable. Started investing recently when the market prices were a bit high,today I am more than 60% down!
Don’t be confuse buying the dip in a bear market, with guaranteed future returns. Just because that company is down 60%+ from ATH does NOT make it a sound long-term investment. Make sure you’re investing in great companies. kudos to Devion Jane
I agree just reached my goal of $500k yearly trade earnings. Setting realistic goals is an essential part of trading..
Please educate me, i'm willing to make consultations to improve my situation,
Her technical analysis is excellent and her interpretation/projections of the market are so accurate I sometimes ask myself if she is a human The point she is the perfect trader to follow for advice and guidance
Wait is this the same lady? That helps me with 17k investment “wow" such a small world..
I think Teslas robotaxis will come faster than most think. The progress is not linial anymore but exponential. More and more compute, more cars and better build software . I think we will see the first this year.
💰
Couldn’t agree more. No one predicted China approval.
Most commentators have no clue as to how advanced Tesla is on self driving. They have never been in a Tesla and certainly never experienced the most recent versions. I discount their comments as being from a ‘no knowledge’ perspective.
Agree but “this year” is way too soon, I would say 3-5. I have used FSD and its amazing, BUT its not there yet.
Everything depends on FSD being ready at some point. Even though I use it a lot, I have no idea when it might be good enough.
I think each city given time, they can farm the data out to make them work. 3 5 10 15 years. I just think on the internet in the 90s vs now.
Re: FSD Perceived as "Good Enough". I've also used it whenever possible, since Mar'22. But the real issue is how much Jonesing do you have when it suspends you for 5 days? I've learned to be MUCH move avoidant of letting that happen, especially since End2End Neural Networked V12 installed! Got V12.3.6 yesterday & can't wait to try autopark/summon! I'm thinking that V12.4.* will be another level up.
As soon as Tesla takes over coverage for FSD topics, it should fly
It’s currently hockeysticking so within the next 24 months FSD will be perfect (I.e. much safer than top 1 percentile of human drivers).
In Toronto Canada and Vancouver Canada, for example, it still has work to do, when comparing other states in the USA. I try not to get my hopes up but I presume it'll be truly dialed in by mid to late 2025. IMHO
Thanks!
As FSD 12 gets closer to 13 (a teenager now) it is at a point where as an interim fix such as Cruise's off site human monitor. With a single intervention in several drives a single monitor could handle multiple vehicles, receive a notification when a questionable situation arose, and handle the crisis. With one intervention per hour and remedying the situation in 3 minutes or less a single monitor could handle 20 cars at a minimum, so allocate 1/15th of a salary for each car, and that worker is easily covered. That covered number of cars would grow with each advancement of FSD until there was no longer a need for monitoring.
As a software/hardware developer of robotic type systems - multiple sensors (lidar radar vision) can create more problems than it solves. Multiple signals need hugely more compute, to solve via AI. Humans use (mostly) vision to drive...imagine if we had a radar screen, and a lidar screen to look at, as well as our vision, when we drive.`First principles thinking says that isn't a good system. Too many inputs can be worse than a top level, single input, (vision) , which is how we blood-bags drive....adding multiple cameras solves the "additional sensor" issue, in my humble opinion....
My take is that multiple sensors types are not inherently bad and the training and inference compute is not that much more - but it is the quality of the sensors is of massive importance. With radar and lidar Tesla was getting massive amounts of false positives. The lidar and rader would see things that are not there asking the driver to take over. Cameras don't really have that issue.
Re: Vision-Only FSD "Sufficient"? I've been following ADS, since the old DARPA Challenge days, from which I concluded that LiDAR lacked sufficient resolution to complete with vision. What's your engineering take on that? I also agree with Munro, that FLIR might be a significant help in highly inclement/dusty conditions. Would FLIR's cost be justified to accommodate those highly rare scenarios or would vision just reducing speed, when temporarily obscured, suffice?
Meanwhile, Re: FSD Perceived as "Good Enough". I've also used it whenever possible, since Mar'22. But the real issue is how much Jonesing do you have when it suspends you for 5 days? I've learned to be MUCH move avoidant of letting that happen, especially since End2End Neural Networked V12 installed from the old heuristic coded V11! Got V12.3.6 yesterday & can't wait to try autopark/summon! I'm thinking that V12.4.* will be another level up, as it seems that NN "adaptations" may actually have started to deploy, mid-revision. I've "detected" such corrections running past the same "troublesome" points in my drives. Is that what you're seeing in V12, vs V11?
This lines up with the “combinatorial explosion” problem Jon Vervake has discussed in AI research. The fewer sensor modalities means greater focus concentrated on one.
Also if it is too foggy or raining to hard or blizzard it white our conditions or in a hurricane or tornadoes it may be best not to drive the autonomous BEV (Tesla RobiTaxi or CyberCab). I was thinking using otter sensors when needed might help but you probably don’t want to be around humans driving in those extreme weather conditions. Thoughts?
Well you humble opinion is mostly correct. The problem can become a "three body problem", impossible to solve, in fact we use filters to avoid some of the problems.
Misleading Headline. Its a suggestion from Tasha, that Robotaxi will first be in California.
People have become so desensitized to lying and deception because we are surrounded by lies, liars, and misinformation and it creeps in at many little ways.
At 9k hours in a year and at $1.00/mile, 50 miles/hour, utilization rate 70%=$315,000.00 revenues minus cost/year=AMAZING! At just 40% utilization: $180k revenues minus costs.
Interesting
So many smart people in the Tesla community
It's a great community for sure
I live deep in the woods of western Massachusetts, our roads are terrible. FSD does great on our no line roads. It does not do well with our potholes, you can easily total your car if you’re not paying attention.
it’s like finding large puddles at night… 💥
Same here in Pennsylvania.
Wait.
Larry and he’s years long experience in business is gold!! 👏🏽👏🏽👏🏽🙏🏼
There has to be an additional $4.00 surcharge per trip on top of $1:00 per mile Tasha’s estimate to cover short trips. Assuming 15 trip per day = $60 per day for surcharges alone. This amount should cover tesla’s monthly payments including the FSD subscriptions , insurance , maintenance and electric charges leaving charges for miles driven as pure profit.
Pre-covid inflation Elon stated that Tesla could run a complete (all costs) robotaxi service for $0.18 per mile. Obviously things like waiting times and repositioning while empty would have been factored in. Bump that up to $0.23 per mile to allow for the approximately 25% inflation bite we've suffered.
Short term Tesla will be able to charge $2 or so per mile. Just make it cheaper than Uber/Lyft plus average tip. Once there are enough robos to replace driven taxis then the rate is likely to start coming down. The magic charge per mile is about $0.60 per mile which makes it cost about the same as driving one's own car.
Tesla, in order to fulfill their mission statement, needs to price private ICEV ownership out of the market.
Selling for $0.60 with a cost of $0.23 is a 160% profit.
100k a year is 274 miles a day. Depending on charging infrastructure (auto charging) that seems a bit high. For longer trips it likely a car can hit that, but a place like SF being 7 miles by 7 miles, even when the car is carrying someone, it will mostly be sitting in slow traffic or stop lights.
274 miles a day is ( 10 trips of 27.4 miles ) so 10 single trips to the airport - very achievable.
Tesla has communicated that the gross profit per robotaxi would be 30 000 usd per year. So that's the number I use but I guess it could be higher at least from the start.
21:44 the log e of 10 / 4 years is 60+ percent per year compound annual gross rate. Please use cagr for growth rates
I'd be very curious how Tesla intends to prevent Chinese government or Chinese competitors from accessing and copying FSD model weights.
Just like in preventing competitors from optically copying the z80 chip: crap trapping information, that looks legit, is included.
That’s why he did it…. Why do u think Tesla factory is actually owned by Tesla US…
Because in return for his nod of approval. He hands over the keys to how things work and how he lays it out
Elon technically came in and gave China an actual right to say they are a competing auto industry. Elon did that…. Single handily…
There is so much focus on August 8th for FSD which I am excited for but there will be an earnings call in July before that how will they handle that without stealing the Thunder from August 8th.
I subscribed to FSD for my new model Y.
What I got is V11.3
Just ensure your car is connected to a Wi-Fi network and you should get the latest update within an hour or two. Alternatively log a service call.
v11.3 is the UI/UX version. there is a different versioning for FSD though it can be confusing at times! You can check what FSD version you have in the software section of settings in your car.
@@HuckWeed Thank you.
@@HuckWeed Huck it says Full Self-Driving (Bata) V11.4.9
I think the current nueral-net version is 12.3
I remember Larry's prediction a month ago about the share price post Q1 earnings, and he thought minimum 10% dip :D
She dodged the questions on how the vehicles are charged. 100k miles per year is 273 miles per day or 1.5 charges. Who, how, and where are you going to charge them? With Tesla slowing down supercharger deployment especially.
night
The range of a model Y is 260. I can see them adjusting the batteries to last one full day. They will charge overnight at a Tesla service station where someone will also clean them. That's not hard.
1.5 charges is 1.5 hours tops. Should be manageable.
They'll be charged by remote, by same group of people who remotely monitor the auto-bots and get them out jambs that they will occasionally find themselves in.
They're going to need a depot to park and maintain them at in each major territory. I would imagine the chargers would be there. That's one charge at night, one partial charge in the day at about 20-25 minutes.
The margins on the robotaxi will be high, but the cost of the fast growing datacenters needed for FSD will also be high. I wonder how these costs are applied in the very positive financial modelspoor.
Regulatory problems are not related to self-driving, rather to local city/county laws related to cab services. You see that happening now in California where RoboTaxi's has to comply with local laws. I'm not saying that can't be solved, maybe using Grok to know local laws, but it's not just as simple as no regulation. There actually will be a myriade of local regulation.
There will always be regions of strong incumbent resistance to new, better ways to do business. Tale as old as time. USUALLY the better economic model triumphs. The bigger the disparity, the better the and faster the conquest...and the disparity here appears pretty huge.
@@737smartin 100% agree, but it does not change the fact that RoboTaxis will have to comply with local zoning laws, in the same way as human taxis. This will require local knowledge to operate, which could come from Grok.
Tasha - solid analysis therefore huge respect. Cathy 2.0 she will be.
I think Monte Carlo is most useful when you have some real data from a small data set, or a small amount of data from a large data set, but not enough to make statistically significant conclusions. It allows you to extrapolate when you don't yet have statistical significance.
If you have no real data at all, and are just making everything up based on guesses, Monte Carlo just gives you more guesses. It does nothing to improve the quality of your guesses. Maybe it allows you to explore how sensitive outcomes are to changes in certain variables, but that's it.
I don't think it makes Ark forecasts any more dependable.
Exactly: bs in = bs out
Florida has already approved autonomous cars.
People are only looking at the supply side to calculate the possible revenue/year. they forget that if all Tesla’s are offering taxi services their will not be enough demand for that many Tesla’s.
Regulators do not intentionally try to prevent companies from releasing products and services, or else there will be no use for regulators. People think erroneously that Tesla’s path to regulatory approval is like going through the eye of a needle.
🚀💰🚀💰 *RoboRickshaw*
I see fsd being solved in the next 12 months, but it will be a few years before robotaxi takes off. Jmo. The average person is pretty skeptical about autonomous driving, they will have to see it over and over before they are convinced it’s safe.
I was in SF a couple weeks back. There were a lot of Waymo robos driving around. I doubt there will be a shortage of willing passengers starting week one. The skeptics can come around later. There will not be enough robos to service everyone at first.
To be approved as a taxi driver FSD does not need to be far better than any human driver, but just equal to any licensed taxi driver. The only caveat is that they have insurance equal to or better than taxi vehicles.
Uber does not gross $140K per car per year! You are mistaken.
Great commentary. On the fence with the new glasses. Ha
Na, clearly on the No side of the fence. She seems to putting in an effort to look more geeky and less attractive
Absolutely.@@oliverbergfeld1606
I believe BYD will partner with Tesla and that's the reason FSD got approved in China. Tesla needs a strong EV partner to take the entire marked rapidly
Supply and demand, exactly. They try to pump the stock, that's all.
FSD and Robotaxi are a dream come true for China. What better way to potentially track and monitor peoples' movements?
As a separate safety issue, FSD cars will probably require at least a mannequin be in the usual driver seat. Other drivers and pedestrians might become distracted when seeing no driver and very likely cause accidents.
That's just people being stupid. You can't fix stupid. I get what you're saying but that's just in the beginning.
What? Like the inflatable pilot in Airplane movie? Haha, that would be funny, but completely stupid too lol
@@didier_777 Yeah, I believe some people are crazy enough to jump in front of a slow moving FSD car just to try and collect insurance. Same thing with other drivers who might quick stop in front of a FSD car just to try and collect insurance. In a crazy world people do crazy things.
@@Mrewink5 will be the exceptions not the rule
@@Mrewink5 @Mrewink5 The difference is that if anyone does that there are 8 cameras that will record everything so I don't think people will be dumb enough to try to scam the insurance companies with video footage. If anything those people will stay away from Teslas.
Yes but also they’re ignoring demand and whether it will be there and constant
Tesla already has a robot on wheels, just not deployed in factory...yet.
Monte Carlo method is great, averages out range of guesses of different inputs.
Intended for realistic confidence interval.
Need sensitivity analysis too to determine influential factors in model.
Beware of GIGO garbage in, garbage out.
These tools should only be used by those with high awareness of what they are using. Think chainsaws.
To increase TAM, need to aim for much lower rates than Uber.
Tesla TaaS won’t merely replace Uber / taxis. That’s tiny market.
It’s got to replace vehicle ownership, even substitute for public transport.
Big challenge is to get to shared travel - almost always we travel like a herd, same direction.
That won’t be on re-purposed private vehicle.
Next generation can’t come soon enough, dedicated taxi design, for future, not past.
Obviousy you don't live in a city. Robotaxi can't substitute for public traffic. If everyone that took the subway in New York City took a Robotaxi instead to work I don't think that many would make it. That would be one increadible traffic jam.
@@arseliopereira2541 trains only work well as you describe, are poor service everywhere else.
When younger, I found a pushbike to be quicker.
Bus is poor everywhere in my experience. Trams are poor compromise.
But it’ll be a challenge to get TaaS to work as I propose. Who’s up for it?
@@arseliopereira2541 It will definitely replace buses but maybe not trains.
@@iandavies4853 I'm not suggesting that it's better to ride a bus than in a Robotaxi. I'm just saying that there isn't enough road capacity for eveyone to drive into the downtown of most cities in their own vehicle.
@@arseliopereira2541 exactly. Hence need for shared vehicle, not single occupancy.
I see lots of huge 50 seat buses with 2 passengers (off-peak).
Still same convoluted route, picking up no-one.
CBDs are different problem.
We all talk to our own world. That’s not mine.
Googled fsd for Canada...at least one province has outrightly banned it."Earlier this month, changes to B.C.'s Motor Vehicle Act quietly came into effect, prohibiting the use of fully automated self-driving vehicles.Apr 22, 2024"
Robotaxi prices won't be super cheap simply because costs are less. They will charge whatever the demand allows, just as they did with the Foundation Series gimmick.
All they have to do is undercut Uber, they don't have to cut the price in half.
In fact they can charge the same price because people, especially women, will choose the Robotaxis to avoid dealing with an unknown human being.
True value (productivity gain) requires change in behaviour, so RoboTaxi will have much more impact in low income countries (China, India), where people does not own a car, and with RoboTaxis can afford to drive. In high income countries (USA, Europe) where people already have a car.
❗️❗️❗️❗️❗️
As much as I like Herbert and I really do, I cringe just about every time he shares time with Larry because he keeps cutting in, doesn’t let Larry finish his sentence and/or thoughts, never mind expanding on his (Larry’s) thoughts.
Larry at 78, has a lot of experience but still a young and furtive mind albeit with the restraint of innumerable years in business experience.
This guy, Larry, is just such a resource of knowledge and measured thoughts of future outcomes.
ROBO-TAXIS ARE INEVITABLE - Just so much better than human drivers. FSD 12.3.4 is amazing, not 100% all the time but very close. 8/8 will be interesting!!
It will need time to charge
Do you think everyone forgot about charging?
100,000 per year is 270 miles per day, so one charge per day for around 35 minutes.
And autonomous can't be available 24/7 because it has to charge. It'll be available 22/7.
I'm glad I'm not the only one who was SHOCKED by not only China's acceptance of FSD,
but also the impact it had on putting TSLA back on the Wall Street track. Thanks Larry and Herbert!
When it came to Tesla as a private robotaxi:
did Larry Goldberg only think once about
the fact that it requires interior and exterior
maintenance after every single ride?
At least the essentials?
Just like every taxi driver can only survive if
his vehicle is tip-top and in very good condition?
Otherwise no one would get into someone
else's car that wasn't absolutely clean.
Tesla will launch an actual robo taxi service on 8-8. IMO.. Austin/Vegas/San Francisco.
Uber drivers do between 140 and 200 per day. At 100,000 miles per year, the Robotaxi is driving 274 miles per day. It doesn't account for any downtime for repairs or other maintenance that need to be done for a car driving that much.
The car needs to go to the nearest Service center and good luck waiting for parts to be there.
Overly optimistic models by Tasha without explaining any detail.
Better analysis is done by Cern Basher and has a ton of realistic assumptions.
Uber drivers don’t make 140k per year - I know this for sure. Average is probably around 50-70k
They mean total revenue not just the driver.
Chances are Musk is overestimating the pace that FSD will be adopted, assuming it reaches level 4/5. $100k is the same price that NY charged for taxi medallions. At the peak there were 300,000 taxis in the US. Very few people will dish out $100k. Tesla needs a sub $25k car for the masses.
She said "double digit thousands" per car per year, not Herbert's "hundred thousand dollars per car per year".
The robotaxi/cybercab doesn't even need to be anything different than what they are already producing. Just being able to summon a regular model 3 or Y would be better than UBER. I don't need the driver. I can drive myself. What I need is for a car to show up at my door when I want it. Using "summon" through the Tesla app I could have any one of the 6 million Teslas currently on the road if they are put into the "taxi fleet". The future "cybercabs" in the densly populated cities don't need a dash or steering wheel or any of the instrument cluster. Think about that.... You could instead have a large compartment with rear facing seats where the dash/steering wheel is currently located and the front facing rear seat in the back...amazing leg room, seats 6 VERY comfortably in a car the size of a VW Golf.
I Cant wait until this is widespread
I assume she factors tires, and battery costs. Warranty gone after one year. Who knows how insurance would work as there is no driver.
is that an apple samsung…. divide..
Double digit profit per car per year is 10k-99k i believe
Why would Tesla license FSD? Small ROI compared with owning.
Let’s not be daft….not more than 1/2 of all buyers will want their car shared !!! It will take years to get to that number too !!!
Elon is going to make 10m Robotaxis a year for exactly that reason.
I would want my robotaxi to work for me.
Hm, in the US maybe, in Europe I doubt ppl will not share. Plus it is likely that TSLA will provide a rating system like Airb&b.
Apart from that, would you not buy a car that would pay for its self, even is it takes 2, 3 or 5 years..?
@@heythave yes and what..are not the average BEV a punter of the future ? Remember the problem with the market now isn’t just interest rates it’s that we’re no longer in the market movers stage but the “laggards stage” with BEV’s ! How many of your acquaintances say they hate Elon and won’t buy a car from him ?
They will be making cars to sell and others to keep in their fleet. At first maybe it's one of 20 cars that will be robotaxi and at any point they can just ramp up as demand increases.
25:36 Herbert, you frequently talk over people; please sort out your lag issues!
An issue (maybe temporary) could be pushback by Taxi drivers/Uber drivers. It's a lot of jobs which could disappear, a very old industry. Consider how much taxi drivers battled against Uber. Now consider Taxi's AND Uber against robo-taxi?
People wonder how Tesla could be worth a significant portion of worlds GDP. That's what happens when you convert unpaid labor (every unpaid driver) and convert that into revenue
How does $1 per mile turn into $100k per year? The math doesn’t seem to make sense. They expect 100k paid miles per year?
LOL!!
When August 8 they will be pushing ROBO Taxi out on stage just like they did the Tesla Bot that could not walk!
The narrative should soon change to "how many lives will be SAVED by FSD!". Insurance liability will differentiate what is safer.
Who lost money investing in ARKK?????
For fun 100 Million Robo-Teslas AT 70% UTILIZATION * $315,000.00=3.15 TRILLIONS in revenues/year at 85% profit margin=$2.67 TRILLION IN PROFITS!!!
Electric robo taxi has to charge. It can't be available 24/7.
This lady also said Tesla should release their Robotaxis back in 2019. Shes been bullishly wrong b4
Why is nobody addressing the demand side ? US have like 1.5 million uber cars. And even if tesla take like 1 million cars, even then it would not make much profits.
Wouldn’t we look at demand for ground passenger transportation at many price points? Not just a direct comparison to a single company.
Well one big factor in the demand side is the cost. Electric miles are cheaper. Autonomous miles are cheaper. Combined they will cut Uber out of the market. But the big change will be people opting to use Transport As A Service rather than buying a car.
The profits per ride are much higher for a Tesla Robot taxi. Uber will disappear immediately where Tesla competes
Uber is expensive af. If robotaxi could take me to work at the same cost I pay in gas but still less than an Uber I’d pay it so I can chill and read while it drives me
@@codingispower1816 i agree. But for that the cost of robotaxi has to be much below 1 dollar per mile. Also it might take people a lot of time to adjust to driverless cars. Also there might behuge resistance from current uber drivers and government because of huge loss of employment.
if it does reach 2000 in 2027 i ll retire so i hope so 🙂
Sure, Tesla is always in the future but never quite gets there LOL.
Autonomous taxi companies and investors are purposely pushing for higher taxi driver wages to squeeze human labor competitors out of the region by making operation costs greater then profitability.
Once they force a minimum wage increase (under the guise of higher labor standards for drivers) they force a region out of the market. This shows then to have on hand a solution, their autonomous products. And for nearly the same cost as a human driver these robotaxis can be deployed in a region saving clients money, reducing co2, relieving congested roads, safety of ai over human error.
Sketchy conspiracy and easy way to force ai over human by increasing labor cost at the government level.
I hope robotaxi has capability of being tele operated. Self sufficiency in 100.0% will be, I suspect, insanely harder than even 99.99%.
惊喜Tesla 將推出Cyber系列大气霸道車款KO全球车企:配套1:星鏈接所有Wi-Fi,配套2:车体太陽儲能全部用电;配套3:FSD腦神經意念控制;配套4:BOT陪伴隨心所欲……
Well, unless Elon lays off the entire Robotaxi team next week.
Its fine, they will be replaced by Optimuses
labor will be replaced by robots
Haha. He is busy hiring over 4000 people for the Robotaxi project.
If some need to go, let them go
@@SupaChinido Elon will be laying off the Robots next month.
Skeptic conspiracy theory:
Talking to uber lyft fulltime drivers, they hadn't really tried to raise minimum wage, many were fine as they were. Now in the twin cities MN there's a push to raise taxi driver wages, and this has caused these services to plan on leaving the region. So if it's not the taxi drivers urging wage increase who is?
Possibly the proponents push for higher taxi wages, are the investors of these AI robotaxi services? Who's to gain the most from current taxi services to withdraw from a region when costs are greater then profit?
A competitor that wants to lock in a profit margin while squeezing out the human labor competition. Because no one invested in a company that only breaks even. Ethics and morals and client savings are just advertisement for greedy investors.
It does seem very sketchy that at the same time taxi wages are being debated, an autonomous competitor would be ready to deploy with in the same time frame as services withdraw from a region. I am concerned there's some dark dealings our politicians are dabbling in for personal profits.
Dear Tesla, keep ramping up Cybertruck, Go pedal to the metal on new lines for CyberCab, Keep the first 1M you produce and start up the Fleet, Never stop building and take care of your employees. Your the one baking the pie, only normal that you get to eat it too.
Tesla’s in good condition, but do you really think they could afford to pay for 1m vehicles in one go?
I’m sure they’ll be selling in excess of 50% of the cars off the line.
Bear in mind the risk of building monopolies - Tesla won’t want to be faced with action on that. The pie is vast - one company doesn’t need to have it all.
What's the truth?!?! Tesla ALSO fired the destination charger team. Hotels and condos IN THE MIDDLE of installs are on hold, the work stopped, and emails to Tesla are bouncing!!!!
Is Musk ok?
Tesla bot - seems to not be factored in?
My respect for Tasha's analysis hasn't wavered over the past 3 years. But I preferred her appearance BEFORE she transitioned.
She doesnt believe too😅
So ARK expects Tesla to be worth $2k a share in a year?
by 2027
@@maxs351 Thanks
2027 is not next year.
2035
Invite Tasha instead of the same 3 random people and only friends you got
Ive interviewed Tasha and many other institutional analysts already. These videos are in the library. I interview new guests regularly including CEOs on a weekly basis.
Supreme Court just ruled that Musk MUST get attorney approval to tweet about Tesla!!! Lets go SEC! Cook this crook! Enough of his lies.
GFY
Too much analysis sometimes All in all, Regulations are a non issue
And what prevents government and unions from featherbedding and passing laws which require driver in the who may or may not do anything but add cost? Musk and the public are not facing the fact that Musk is smaller than the government. His only option will be do what has been done for ever all over the world and that is to pay off politicians. I suspect he would walk away before dealing with that.
Ok. First the video talks about Tesla takes most.
Then the video talks about why China invited Tesla and their FSD into China.
I can see the conflict between these two views, can you?
It's a big lie. I would never rent my car out to anyone! Your insurance company won't insure your car! Also who trust a self driving car!
*Tesla going to 0$ elon musk will take the fsd to xai because tesla shareholders won't give elon 50 billion* 😂
If you can get $100k per year per car, why bother selling the cars at all? Just keep them all, shutdown the car sales business, and become a taxi company.
I stoped watching some time ago. Today you had larry on and this told me to stay off.
Just a philosophical question. Why should the citizens of America treat Elon's robotaxis with respect when he shows such disrespect for so many of America's citizens? (including employees)
Why should anybody pay any attention to your nonsense You're offensive j
The most important task for a CEO is to guide the company to prosperity. Not to keep his employees at all costs
He also employs 140,000 people, minus the 500 he just fired.
Moronic take.
We shouldn't show them respect and we will fuck with them to no end. Not to mention the number of highschoolers who will endlessly troll autonomous anything. It is gonna be a TH-cam clip gold mine.
$100K operating profit per year, right!!!!!! Perhaps before buying into this bullshit ask Hertz their opinion about Tesla.
Hertz also operates as a pseudo car dealer. Buying from the OEM,s at large discounts. After a short while they sell them for maybe a profit or a small loss, large part of their business model. With Tesla dropping prices to make them more affordable for the masses, Hertz took the decision to sell as the value for their business model was dropping