Is Tesla a good investment?

[Fru-Gal]

I love technology and don’t doubt that when well designed it can be miraculous.

I’m not comparing vision+neural net to a human. I’m comparing it to the competition.

End-to-end learning is also problematic. I guess it is characteristic of a Musk company that they would double down on a singular (& cheap) solution rather than try to ensure success by making a robust solution that selects the best-of-breed technologies used by other transportation modalities.

I also, as an investor, wouldn’t believe any company that claims to have an AI moat. AI is becoming commoditized.

By your logic, Ford is cutting corners on safety because it doesn't offer lidar as a driver assist/safety feature on its current cars. More is always better right?

Again, if humans are deemed safe enough to drive using our brains (neural net) and and eyeballs (vision) without the addition of Lidar, why would we deem it necessary for a superior neural net with superior vision (9 cameras with faster reaction time) to add lidar?

We are constantly, as a society, triangulating between cost, risk, and safety. Is it worth 10X the cost for a nebulous gain in safety, when Tesla's FSD brings cost effective autonomous driving to the masses and dramatically reduces the number of fatalities on the road each year? Widespread adoption is what is needed to drive down fatalities, not added cost for little to no actual safety gains.

It’s notable that you don’t address the specific points I made. Instead of comparing future Teslas to Ford’s existing human-piloted cars, can you compare them to Waymo’s?

Or, how about comparing FSD to other companies’ ADAS:

https://www.forbes.com/sites/brookecrothers/2024/12/29/tesla-full-self-driving-vs-the-rest-be-very-careful-choosing-your-next-ev/

This is more evidence that any self-driving moat is fast disappearing.

Also, what is your evidence that the FSD neural net is equal or superior to a human brain? Does Tesla publicly test with things like https://github.com/carla-simulator/leaderboard/blob/master/README.md

How confident are you in end-to-end reinforcement learning (the Tesla and Wayze approach), including for testing, scalability and extensibility?

[neo von retorch]

What was the investment to R&D, build tooling, and produce the Cybertruck, and what is their current ROI?

What are the assumptions on Cybercab ROI?

The average cost of an Uber ride in the US is $2/mile. Let's assume a Cybercab undercuts this at $1.50/mile average. A Cybercab can operate 16 hours a day, leaving 8 hrs for recharging, maintenance and down time. Let's assume an average speed of 35 mph. So,  16hrs x 35 mph = 560 miles/day x 7days/week x 52 weeks = 203k miles/yr. Let's be super conservative and assume suboptimal utilization and each Cybercab does 100k billable miles per year. At $1.50/mile that's $150k in revenue per Cybercab per year. Fuel cost is roughly $0.15 per KwH. Cybercab gets 6 miles per KwH. So, annual fuel cost is $2,500 (assuming Tesla does provide their own energy via solar). Upkeep on a Cybercab is minimal. Cordless charging pads and robotic car cleaners take care of refueling and cleaning while charging. Maintenance cost is near zero. No scheduled maintenance in year one for a Tesla. Maybe some wiper fluid and one set of new tires. Let's put fueling, cleaning, registration fees and maintenance cost in year one at $5,000. Feel free to double that to $10k, it doesn't matter.

So, a Cybercab with a production cost of $15k rolls off the line and is accepting passengers the next day in Austin. In year 1 it generates $150k in revenue and costs $5k to operate, netting $145k for Tesla in year one. A nearly 10X ROI in the first year. The cab would pay for itself in under 2 months and produce enough profit in year 1 to build 6 additional Cybercabs.

Would like better sources, but claude.ai says that 30-40% of taxi miles driven are paid / with passengers and 60-70% are between fares. Some of those reasons wouldn't apply if there's no human, though many would:

Driving to and from initial pickup locations
Circling or waiting for fares
Returning to taxi stands
Repositioning between potential high-demand areas
Personal or maintenance-related driving

Obviously "returning to charging port" would be on there. Human taxi drivers seem to drive 45-70k, though depending on charge cycles, as well as peak and off-peak hours, it's fine to estimate 100k total miles being driven by an autonomous electric taxi.

Now assuming 50% of those are paid, because the machines are just better at algorithms than the existing humans + dispatch + software, you're looking at 100k miles, 50k in fares. Estimates for economy taxi rides in Austin, Tx are already around $1.50 / mile, so perhaps they undercut it at $1.25 / mile? And the only cost is $0.15 / mile electric charging. So over 100k miles, that's $15,000 in fuel cost, and $62,500 in revenue. (And as mentioned above, $5-10k in "other maintenance", hand-wavy napkin math!) If all of that checks out, it's still close to $40k / year profit per vehicle if all the assumptions are correct.

Here's a question though. Why are you questioning the economics of the Cybercab on an investor forum and asking others to do research that you could have done on your own before posting shade about the product?

Are the only people that can question if TSLA is a good investment, the people that are for investing? How about some contrary opinions so we can think about the things we haven't already? That goes both ways. If you're against it, why aren't you allowed to ask "why am I wrong, how could I be convinced this would be worth my investment dollars?"

[bacchi]

Everyone keeps mentioning Tesla approval for the Cybercab in Austin. Is there a link to this regulatory approval? Is it like California's approval, which is employees only with a driver?

Texas amended its transportation code in 2017 to allow autonomous vehicles to operate on its roads, and it took away any ability for local governments to restrict testing or deployment.

Texas' rules say that autonomous vehicles are allowed to operate throughout the state as long as they comply with state traffic and vehicle laws and conform to federal rules regarding vehicle safety and AVs

  - https://arstechnica.com/cars/2025/02/tesla-turns-to-texas-to-test-its-autonomous-cybercab/

Ok, so there's no approval process.

Tesla is hiring teleoperators. That's...good I guess?

https://www.tesla.com/careers/search/job/c-software-engineer-teleoperation-tesla-bot-and-robotaxi-227959

"As we iterate on the AI that powers them, we need the ability to access and control them remotely. "

[ColoradoTribe]

Relying exclusively on vision and “neural nets” is a laughable strategy.

Would you fly in a commercial jet if reassured that it was being piloted with a vision system and neural nets exclusively?

Let’s think about things that navigate beyond cars. Things like defense systems, submarines, planes, trains you need multiple instruments and sensors in order to ascertain your surroundings when visibility is problematic, or there are other difficult conditions. once again musk is hamstrung by his hubris. He has created a brilliant infallible system — that does not even work right now. Meanwhile, Waymo is using lidar, radar, vision and human oversight — and has several years of real world experience.

I do think the idea that every individual Tesla would become a cyber cab for the owner is a compelling thought experiment. But I don’t think there’s any proof that that can happen as the cars are currently instrumented.

Have you ridden in a Tesla operating under FSD supervised? I have.

How do you drive a car? You use vision (eyeballs) and a neural net (brain) to operate the car. Same as Tesla, except the AI is never emotional, or tired, or drunk, or distracted, or sleepy, or get s a leg cramp. The neural net also has the advantage of billions of miles of collective driving experience and 9 cameras (or nine sets of eyeballs) constantly surveying blind spots and approaching traffic. And reaction time superior to any human.

Given how many examples there are in the world where a machine/computer is superior to a human at performing a task (greater speed, accuracy, precision, etc), its seems like hubris or naiveté to believe a human's ability to drive a car can never be surpassed. In fact, the current version of FSD is already as good as an average human driver under most conditions and is improving exponentially.

What do you think the realistic revenue potential is for this robotaxi service, and what are realistic margins?

I did the math further up-thread, but a highly optimistic scenario seems to be a $10B - $15B/yr business in a decade.  However, it would most likely be significantly lower than this.  For context, Tesla already has about $100B in revenue.

It also isn't an 80%+ margin AI company.  It's not even a 30-50% margin business.  It's a capital inefficient business that can probably get to 20% gross margins initially, which would go down to the 10% range if they try to price below human-operated cars.

It's combining the capital inefficiency of a car manufacturer with the capital inefficiency of Zipcar or Hertz.  Except Zipcar has a secondary market for their cars that a robotaxi won't. 

It may turn into a worthwhile business venture (I'm skeptical of this, but don't rule it out), but the idea that it should command some valuation premium is pretty absurd.

Provided my math and assumptions for year 1 revenue and ROI for a Tesla Cybercab above. Feel free to extrapolate from there. Tesla electric drive train and battery pack are capable of 500k miles or more as demonstrated by early Model 3s that have surpassed this milestone and more. This puts the life span of a Cybercab at 5 plus years assuming 100k miles/yr, which would give lifetime revenue for a single Cybercab around $750k. Minus maintenance and fuel cost of $5k/yr and production cost of $15K, so around $700k profit per Cybercab.

[ColoradoTribe]

I love technology and don’t doubt that when well designed it can be miraculous.

I’m not comparing vision+neural net to a human. I’m comparing it to the competition.

End-to-end learning is also problematic. I guess it is characteristic of a Musk company that they would double down on a singular (& cheap) solution rather than try to ensure success by making a robust solution that selects the best-of-breed technologies used by other transportation modalities.

I also, as an investor, wouldn’t believe any company that claims to have an AI moat. AI is becoming commoditized.

By your logic, Ford is cutting corners on safety because it doesn't offer lidar as a driver assist/safety feature on its current cars. More is always better right?

Again, if humans are deemed safe enough to drive using our brains (neural net) and and eyeballs (vision) without the addition of Lidar, why would we deem it necessary for a superior neural net with superior vision (9 cameras with faster reaction time) to add lidar?

We are constantly, as a society, triangulating between cost, risk, and safety. Is it worth 10X the cost for a nebulous gain in safety, when Tesla's FSD brings cost effective autonomous driving to the masses and dramatically reduces the number of fatalities on the road each year? Widespread adoption is what is needed to drive down fatalities, not added cost for little to no actual safety gains.

It’s notable that you don’t address the specific points I made. Instead of comparing future Teslas to Ford’s existing human-piloted cars, can you compare them to Waymo’s?

Or, how about comparing FSD to other companies’ ADAS:

https://www.forbes.com/sites/brookecrothers/2024/12/29/tesla-full-self-driving-vs-the-rest-be-very-careful-choosing-your-next-ev/

This is more evidence that any self-driving moat is fast disappearing.

Also, what is your evidence that the FSD neural net is equal or superior to a human brain? Does Tesla publicly test with things like https://github.com/carla-simulator/leaderboard/blob/master/README.md

How confident are you in end-to-end reinforcement learning (the Tesla and Wayze approach), including for testing, scalability and extensibility?

Please read up thread before responding further. I addressed the competitiveness of Waymo's model in considerable detail earlier today. Even the CEO of Google, Waymo's parent company has conceded Tesla FSD is in the lead. Quick Google search will find you the video clip.

[reeshau]

It's less a question of if and more a question of when. Once FSD unsupervised gets regulatory approval, no one will buy a car in the US that doesn't come with FSD capability, which means every car company will need to either license Tesla's software or spend billions trying to create their AI capabilities and a data set of billions of miles driven to train their own neural net. Regulatory approval is a given once insurance companies and regulators see data showing how many lives and dollars FSD saves.

The day FSD gets approved, it will at best be the second level 3 approval.  Mercedes Drive Pilot was the first.  And there are many more also striving to achieve this.

https://www.mbusa.com/en/owners/manuals/drive-pilot

That isn't to say the first always wins, or the best always wins.  The classic example is VHS / Beta.

There are also safety regulations that may not be achieved by a vision-only system,.   I doubt the 2029 NHTSA pedestrian safety regs will stand under the current administration, but they were chasing Europe.   The nighttime scenarios, in particular, are challenging, as is well-documented not just for Tesla, but other vehicles like the Ford Lightning.

Autonomous driving is an achievement.  Autonomous driving in any weather, which is necessary to provide a reliable service, is another thing on top of that.

[Fru-Gal]

I watched that YT short. Pure politics. All Pinchai says is that Tesla is the leader, Tesla and Waymo are the top two. Also I wouldn’t trust a single tech CEO as far as I could drop kick them down the road. Especially those toadying to presidents Trump/Musk.

I am asking more specifically about what experts in self driving have to say! What are the academics studying, what are the training/testing systems out there now and how does Tesla rate on them?

Still haven’t addressed any of my other questions. Not about the potential or the future of end-to-end reinforcement learning for self-driving, or “neural nets” in general. But about the current capability.

I just posted above an article comparing FSD to other advanced driving systems. Notably, the author said that while FSD was extremely impressive and, unlike all the others, offered highway driving, THEY DID NOT TRUST IT to not occasionally do stupid or weird things.

[reeshau]

Interesting development in China: Tesla has dropped FSD from its branding there.

https://www.teslarati.com/tesla-updates-fsd-branding-china/


Tesla has tweaked the naming of its smart driving system offerings in China, with the company dropping “FSD” terminology from its vehicle order pages. The update was observed by industry watchers earlier this week.

Names Adjusted, Features Intact
Tesla China’s RMB 64,000 ($8,820) package—once listed as “FSD Intelligent Assisted Driving”—has been updated to “Intelligent Assisted Driving.” Its RMB 32,000 mid-tier system, previously dubbed “Enhanced Version Automated Assisted Driving”, has also been updated to “Enhanced Assisted Driving.”

Tesla’s basic Autopilot system, which was previously dubbed “Basic Version Assisted Driving,” has been changed to “Basic Assisted Driving” as well. Even the system’s umbrella term has been updated from “Autopilot Automated Assisted Driving” to simply “Assisted Driving Package.”

[GuitarStv]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

[reeshau]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

Yes, another back office hurdle!

At least if Tesla owns the whole stack: car, software, driving service, it will make it easier to name the defendant.  I bet the ambulance chasers will be drooling at one with such deep pockets.

In all seriousness, this may be a key reason to go slowly.  Unfortunately, it will take a real test case to start to figure it out.

[Telecaster]

I'm not sure what the concern is based on what you are expressing here? Tesla has regulatory approval to operate a Cybercab network in Austin starting in June. How many vehicles is rather irrelevant at this stage. They will use Austin to work out kinks and as proof of concept. In the meantime, a Cybercab production line is underway in the Austin Gigafactory. Once Tesla is satisfied with the Austin trial results, they will expand to other cities and fire up mass production. At this stage, the number of Tesla Cybercabs surpasses the number of Waymos in a matter or weeks.

Tesla is cash flow positive and sitting on billions (16B+ end of Q4). It does not matter that they have to front the cost of each Cybercab ($15-$20k). Those vehicles will pay for themselves in less than a year and then start generating revenue for Tesla to build more Cybercabs (snowball rolling downhill at this point).

Reminder here that every Tesla with HW4 or better also becomes a cybercab for its owner the day unsupervised FSD is approved on a state or federal level. Additionally, unlike Waymo, Tesla doesn't have to GPS map and geofence each new service area. Tesla's vision + neural net FSD model will enable a new cybercab to operate anywhere in the US just like any human driver in the US is capable of flying into a city they've never visited before and operate a car safely.

I don't have an opinion on the economic of the Cybercab, but I think you are overstating the capabilties and understating the difficulty.  We don't have many details about the Austin rollout, but it it almost certainly will be geofenced (it definitely will only be operating in a limited area) and use will remote human monitoring.   That poses one of the same challenges that Waymo has, namely it is difficult to scale.   

Even the Austin experiment is a smash hit, Tesla is only in the beginning phases of the permit process in California and no other locations that I'm aware of.   Picking up passengers for hire is a regulated industry everywhere, and I don't think the regulatory framework for autonomous pickup even exists in most locations, so I'm not at all as optimistic as you that it can expand in a matter of weeks once the tech is developed.   If the tech was available right now, it would still be a matter of years, not weeks.   

Reminder here, that Musk previously promised that every Tesla with HW3 or better could be fully autonomous without hardware updates.   As per Tesla, that promise has been broken.   From an investing standpoint, it would be prudent to be skeptical about other claims in this area without some form of verification.  It is a reasonable question to ask if HW4 will actually be sufficient as promised,  or if it will require a further hardware update.   

Finally, I watched the video you posted and it basically echoed Tesla's position  that most of the recent drop is attributable to the Model Y refresh  (which despite what the video stated has been widely reported in the press).   However, that's not dealing with the elephant in the room.  Namely,  Tesla lost both sales and marketshare in 2024 compared to 2023, even as the global EV market grew by 25%.   

Since this is a stock analysis board, let's do a little analysis:

Net Income
2022: $12.56 billion
2023: $15.0 billion
2024: $7.09 billion

Profit Margin
2022: 15.4%
2023: 15.5%
2024: 7.3%

7.3% margin is still good for a car company, but terrible for a tech company.    Decreasing margins to increase market share isn't necessarily a bad thing, but Tesla isn't increasing sales or market share.   Decreasing margins to maintain sales should be a red flag to investors.   

As an aside, I just checked the Tesla website, and refreshed Model Y's are available for immediate delivery, no backlog.   

[ColoradoTribe]

What was the investment to R&D, build tooling, and produce the Cybertruck, and what is their current ROI?

What are the assumptions on Cybercab ROI?

The average cost of an Uber ride in the US is $2/mile. Let's assume a Cybercab undercuts this at $1.50/mile average. A Cybercab can operate 16 hours a day, leaving 8 hrs for recharging, maintenance and down time. Let's assume an average speed of 35 mph. So,  16hrs x 35 mph = 560 miles/day x 7days/week x 52 weeks = 203k miles/yr. Let's be super conservative and assume suboptimal utilization and each Cybercab does 100k billable miles per year. At $1.50/mile that's $150k in revenue per Cybercab per year. Fuel cost is roughly $0.15 per KwH. Cybercab gets 6 miles per KwH. So, annual fuel cost is $2,500 (assuming Tesla does provide their own energy via solar). Upkeep on a Cybercab is minimal. Cordless charging pads and robotic car cleaners take care of refueling and cleaning while charging. Maintenance cost is near zero. No scheduled maintenance in year one for a Tesla. Maybe some wiper fluid and one set of new tires. Let's put fueling, cleaning, registration fees and maintenance cost in year one at $5,000. Feel free to double that to $10k, it doesn't matter.

So, a Cybercab with a production cost of $15k rolls off the line and is accepting passengers the next day in Austin. In year 1 it generates $150k in revenue and costs $5k to operate, netting $145k for Tesla in year one. A nearly 10X ROI in the first year. The cab would pay for itself in under 2 months and produce enough profit in year 1 to build 6 additional Cybercabs.

Would like better sources, but claude.ai says that 30-40% of taxi miles driven are paid / with passengers and 60-70% are between fares. Some of those reasons wouldn't apply if there's no human, though many would:

Driving to and from initial pickup locations
Circling or waiting for fares
Returning to taxi stands
Repositioning between potential high-demand areas
Personal or maintenance-related driving

Obviously "returning to charging port" would be on there. Human taxi drivers seem to drive 45-70k, though depending on charge cycles, as well as peak and off-peak hours, it's fine to estimate 100k total miles being driven by an autonomous electric taxi.

Now assuming 50% of those are paid, because the machines are just better at algorithms than the existing humans + dispatch + software, you're looking at 100k miles, 50k in fares. Estimates for economy taxi rides in Austin, Tx are already around $1.50 / mile, so perhaps they undercut it at $1.25 / mile? And the only cost is $0.15 / mile electric charging. So over 100k miles, that's $15,000 in fuel cost, and $62,500 in revenue. (And as mentioned above, $5-10k in "other maintenance", hand-wavy napkin math!) If all of that checks out, it's still close to $40k / year profit per vehicle if all the assumptions are correct.

Here's a question though. Why are you questioning the economics of the Cybercab on an investor forum and asking others to do research that you could have done on your own before posting shade about the product?

Are the only people that can question if TSLA is a good investment, the people that are for investing? How about some contrary opinions so we can think about the things we haven't already? That goes both ways. If you're against it, why aren't you allowed to ask "why am I wrong, how could I be convinced this would be worth my investment dollars?"

I'm good with the above, while assuming some of your assumptions are on the low side and conceding my assumptions may be on the high side. Either way it is clear that Cybercab can be profitable in short order and undercut any competition. If Tesla bankrupts ride share services like Uber and Lyft and traditional taxi services, what does that do to utilization rates and billable miles per year per CC?

If the cost of using Tesla AEVs is so cheap as to undercut the cost of car ownership in urban and suburban locations what does that do to utilization rates and billable miles/year? If folks start forgoing car ownership and relying entirely on AEVs for ground transportation?

We need to be careful applying traditional metrics and assumptions to a disruptive technology. The ground is shifting under our feet.

I will give you the benefit of the doubt when it comes to the motivation behind your posts. You may very well be seeking objective understanding and contrarian opinions to debate and be open to an investment either way. Unfortunately, I have been on this thread since the beginning and some of the others who have been on for as long continue to move goal posts, dredge up the "concern of the day", refuse to concede when proven wrong, and whose single presence on this thread is to cheerlead against Tesla and amplify any FUD they happen across. Even though they've said time an again they would never own the stock.

[ColoradoTribe]

Everyone keeps mentioning Tesla approval for the Cybercab in Austin. Is there a link to this regulatory approval? Is it like California's approval, which is employees only with a driver?

Texas amended its transportation code in 2017 to allow autonomous vehicles to operate on its roads, and it took away any ability for local governments to restrict testing or deployment.

Texas' rules say that autonomous vehicles are allowed to operate throughout the state as long as they comply with state traffic and vehicle laws and conform to federal rules regarding vehicle safety and AVs

I imagine, even in a mature AEV market, there will exist a role for teleoperators. What matters is the ratio of teleoperator to AEVs and the rate of needed interventions/rescues. If one teleoperator can efficiently oversee 100 AEVs, does that mean they're not really autonomous? I don't know the current ratio, but I suspect this is where we are headed.



  - https://arstechnica.com/cars/2025/02/tesla-turns-to-texas-to-test-its-autonomous-cybercab/

Ok, so there's no approval process.

Tesla is hiring teleoperators. That's...good I guess?

https://www.tesla.com/careers/search/job/c-software-engineer-teleoperation-tesla-bot-and-robotaxi-227959

"As we iterate on the AI that powers them, we need the ability to access and control them remotely. "

[ColoradoTribe]

I watched that YT short. Pure politics. All Pinchai says is that Tesla is the leader, Tesla and Waymo are the top two. Also I wouldn’t trust a single tech CEO as far as I could drop kick them down the road. Especially those toadying to presidents Trump/Musk.

I am asking more specifically about what experts in self driving have to say! What are the academics studying, what are the training/testing systems out there now and how does Tesla rate on them?

Still haven’t addressed any of my other questions. Not about the potential or the future of end-to-end reinforcement learning for self-driving, or “neural nets” in general. But about the current capability.

I just posted above an article comparing FSD to other advanced driving systems. Notably, the author said that while FSD was extremely impressive and, unlike all the others, offered highway driving, THEY DID NOT TRUST IT to not occasionally do stupid or weird things.

Goodness. You wouldn't trust the word of a tech CEO that puts his own product behind a competitors? That's about the only time I would trust them. Your explanation is he's trying to curry favor with the incoming administration at the risk of demoralizing his own workers and pissing off his investors and shareholders. Is there really any reason to converse further if these are the lengths you're willing to go to avoid giving Tesla any due credit? I know my answer.

Suggestion, why don't you answer some of your own questions and and value my time. I offered several responses today to multiple posters, including you, and provided a lot of substance and citations along the way. Have you answered my questions?

Why doesn't Ford include lidar on its ICE vehicles if its so important a safety feature above the neural net and vision used by both humans and FSD to operate a car?

[ColoradoTribe]

I'm not sure what the concern is based on what you are expressing here? Tesla has regulatory approval to operate a Cybercab network in Austin starting in June. How many vehicles is rather irrelevant at this stage. They will use Austin to work out kinks and as proof of concept. In the meantime, a Cybercab production line is underway in the Austin Gigafactory. Once Tesla is satisfied with the Austin trial results, they will expand to other cities and fire up mass production. At this stage, the number of Tesla Cybercabs surpasses the number of Waymos in a matter or weeks.

Tesla is cash flow positive and sitting on billions (16B+ end of Q4). It does not matter that they have to front the cost of each Cybercab ($15-$20k). Those vehicles will pay for themselves in less than a year and then start generating revenue for Tesla to build more Cybercabs (snowball rolling downhill at this point).

Reminder here that every Tesla with HW4 or better also becomes a cybercab for its owner the day unsupervised FSD is approved on a state or federal level. Additionally, unlike Waymo, Tesla doesn't have to GPS map and geofence each new service area. Tesla's vision + neural net FSD model will enable a new cybercab to operate anywhere in the US just like any human driver in the US is capable of flying into a city they've never visited before and operate a car safely.

I don't have an opinion on the economic of the Cybercab, but I think you are overstating the capabilties and understating the difficulty.  We don't have many details about the Austin rollout, but it it almost certainly will be geofenced (it definitely will only be operating in a limited area) and use will remote human monitoring.   That poses one of the same challenges that Waymo has, namely it is difficult to scale.   

Even the Austin experiment is a smash hit, Tesla is only in the beginning phases of the permit process in California and no other locations that I'm aware of.   Picking up passengers for hire is a regulated industry everywhere, and I don't think the regulatory framework for autonomous pickup even exists in most locations, so I'm not at all as optimistic as you that it can expand in a matter of weeks once the tech is developed.   If the tech was available right now, it would still be a matter of years, not weeks.   

Reminder here, that Musk previously promised that every Tesla with HW3 or better could be fully autonomous without hardware updates.   As per Tesla, that promise has been broken.   From an investing standpoint, it would be prudent to be skeptical about other claims in this area without some form of verification.  It is a reasonable question to ask if HW4 will actually be sufficient as promised,  or if it will require a further hardware update.   

Finally, I watched the video you posted and it basically echoed Tesla's position  that most of the recent drop is attributable to the Model Y refresh  (which despite what the video stated has been widely reported in the press).   However, that's not dealing with the elephant in the room.  Namely,  Tesla lost both sales and marketshare in 2024 compared to 2023, even as the global EV market grew by 25%.   

Since this is a stock analysis board, let's do a little analysis:

Net Income
2022: $12.56 billion
2023: $15.0 billion
2024: $7.09 billion

Profit Margin
2022: 15.4%
2023: 15.5%
2024: 7.3%

7.3% margin is still good for a car company, but terrible for a tech company.    Decreasing margins to increase market share isn't necessarily a bad thing, but Tesla isn't increasing sales or market share.   Decreasing margins to maintain sales should be a red flag to investors.   

As an aside, I just checked the Tesla website, and refreshed Model Y's are available for immediate delivery, no backlog.

Now include RND and Capex from your numbers above to tell the true picture. It's this RND and Capex that is going to fuel the next wave of growth. Those billions spent on Capex and RND is Tesla's competitive advantage going forward and while it eats into current margins and profit, it will field future growth. If Tesla were focused on growing car sales I would be concerned with last year's numbers. However, for better or worse, Tesla is focused on AI, FSD, and Optimus. If successful any one of these revenue streams will dwarf car revenue. Margins and revenue from battery storage products also continue to grow and will eventually equal or exceed car revenue. EV market share will increase with Model Y refresh and introduction of lower cost models that will increase TAM from 5% to 34-40%.

[ColoradoTribe]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

There are roughly 40,000 vehicular deaths a year in the US. This is a low bar for an AEV to clear. Of course we are not going to reduce deaths to zero, but if we can cut that number by 90% why wouldn't we do it? An AI driver will never drive distracted, tired, impaired or recklessly. An AI driver will never have road rage and cause an accident. What insurance company wouldn't insure an AEV if it was 90% less likely to be involved in a fatal accident compared to insuring a human driver? Cost of insuring will fall to the owner of the AEV, but cost of insuring each vehicle will be less than the cost of insuring a human driver and easily covered by revenue generated by a robotaxi.

Specifically, a Tesla operated with autopilot is 6X LESS likely to get in an accident compared to the US average.

https://www.tesla.com/VehicleSafetyReport

When death rates from driving fall in line with death rates from riding public transportation, flying on an airplane, or riding in a train, we will surrender control, just like we do in these other situations ,because we will deem the risk acceptable and better for society at large compared to the alternative of another 40,000 souls a year.

[ColoradoTribe]

We are constantly, as a society, triangulating between cost, risk, and safety. Is it worth 10X the cost for a nebulous gain in safety, when Tesla's FSD brings cost effective autonomous driving to the masses and dramatically reduces the number of fatalities on the road each year? Widespread adoption is what is needed to drive down fatalities, not added cost for little to no actual safety gains.

I am all for this eventually happening.  Safer roads are better, and if driving automation gets us there, great.  But don't put the cart before the horse, we do need to get there first.

From 2021 to the end of 2023, Teslas had 736 crashes, 17 of them fatal, while using automation.  That gives FSD a fatal accident rate of 11.3 deaths per 100 million miles travelled. Humans driving in 2022 had a fatal accident rate of 1.35 deaths per 100 million miles travelled.  Tesla vehicles themselves are the most crashed on the road.  Tesla drivers are involved in 26.67 accidents per 1,000 drivers (more than any other vehicle brand). - (https://www.planetearthandbeyond.co/p/self-driving-cars-are-way-more-dangerous)

The National Highway Traffic Safety Administration's analysis concluded that the Tesla Autopilot death rate is higher than reported estimates. Incidents were caused by the ADAS failing to recognize other vehicles, insufficient Autopilot driver engagement, and (they argue) poor operational design. - (https://www.forbes.com/sites/stevebanker/2025/02/11/tesla-again-has-the-highest-accident-rate-of-any-auto-brand/)

It really doesn't sound like FSD is yet ready for prime time.  Now granted, with Musk running the government it seems unlikely that the NHTSA will be involved in blocking Tesla from doing anything they want in the future . . . but it doesn't paint as rosy a picture as you have been making.

Everyone keeps mentioning Tesla approval for the Cybercab in Austin. Is there a link to this regulatory approval? Is it like California's approval, which is employees only with a driver?

Texas amended its transportation code in 2017 to allow autonomous vehicles to operate on its roads, and it took away any ability for local governments to restrict testing or deployment.

Texas' rules say that autonomous vehicles are allowed to operate throughout the state as long as they comply with state traffic and vehicle laws and conform to federal rules regarding vehicle safety and AVs

  - https://arstechnica.com/cars/2025/02/tesla-turns-to-texas-to-test-its-autonomous-cybercab/

There are opposing stats on this topic. See my post above. I believe the flaw in your stats is the assumed total number of miles driven by Tesla drivers. Tesla actually knows this number and came out and said this death rate assumed a denominator that was way off. There were some very flawed assumptions used to arrive at the total number of miles driven is what I recall. Can you provide a source or the math behind how the total number of miles driven was arrived at?

Teslas are very fast cars with insane acceleration. The fact that there are human caused fatalities is not surprising. If anything it only speaks to the need for FSD unsupervised. An AI driver presumably won't drive above the speed limit or at least maintain a reasonable offset.

[Telecaster]

If Tesla bankrupts ride share services like Uber and Lyft and traditional taxi services, what does that do to utilization rates and billable miles per year per CC?

If the cost of using Tesla AEVs is so cheap as to undercut the cost of car ownership in urban and suburban locations what does that do to utilization rates and billable miles/year? If folks start forgoing car ownership and relying entirely on AEVs for ground transportation?

I'm quite certain Tesla has zero intention of doing this, and the reasons are simple.   If you rely on the AEV model for most public transportation, then there needs to be enough of them to service transportation needs at times of unusually high demand.  Like say, a sporting event.    So if you a cybercab company that means most of your fleet will sit idle most of the time.

It makes way more sense to have enough cabs to service typical baseload volumes so your fleet is busy and occupied all the time, and let Uber and Lyft or whatever service the surge demand times.   That's kind of how the Uber and Lyft model works already.   

Now include RND and Capex from your numbers above to tell the true picture. It's this RND and Capex that is going to fuel the next wave of growth. Those billions spent on Capex and RND is Tesla's competitive advantage going forward and while it eats into current margins and profit, it will field future growth. If Tesla were focused on growing car sales I would be concerned with last year's numbers. However, for better or worse, Tesla is focused on AI, FSD, and Optimus. If successful any one of these revenue streams will dwarf car revenue. Margins and revenue from battery storage products also continue to grow and will eventually equal or exceed car revenue. EV market share will increase with Model Y refresh and introduction of lower cost models that will increase TAM from 5% to 34-40%.

Happy to.  In fact, I looked that up for my prior post, so I've got those numbers handy.   But first, I'm not as confident as you that Tesla can ignore the car business.   AI, autonomous driving, and Optimus are not yet commercial products.   They might be one day, but Tesla has competitors in all those areas, some with deep, deep pockets.   

Capex
2022: $7.2 billion
2023: $8.9 billion
2024: $11.3 billion

So Tesla is definitely increasing investment in this area.  FWIW, Google announced it is spending $75 billion in capex this year, Microsoft is spending $80 billion, and Amazon is spending $75 billion. 

   

[Fru-Gal]

I watched that YT short. Pure politics. All Pinchai says is that Tesla is the leader, Tesla and Waymo are the top two. Also I wouldn’t trust a single tech CEO as far as I could drop kick them down the road. Especially those toadying to presidents Trump/Musk.

I am asking more specifically about what experts in self driving have to say! What are the academics studying, what are the training/testing systems out there now and how does Tesla rate on them?

Still haven’t addressed any of my other questions. Not about the potential or the future of end-to-end reinforcement learning for self-driving, or “neural nets” in general. But about the current capability.

I just posted above an article comparing FSD to other advanced driving systems. Notably, the author said that while FSD was extremely impressive and, unlike all the others, offered highway driving, THEY DID NOT TRUST IT to not occasionally do stupid or weird things.

Goodness. You wouldn't trust the word of a tech CEO that puts his own product behind a competitors? That's about the only time I would trust them. Your explanation is he's trying to curry favor with the incoming administration at the risk of demoralizing his own workers and pissing off his investors and shareholders. Is there really any reason to converse further if these are the lengths you're willing to go to avoid giving Tesla any due credit? I know my answer.

Suggestion, why don't you answer some of your own questions and and value my time. I offered several responses today to multiple posters, including you, and provided a lot of substance and citations along the way. Have you answered my questions?

Why doesn't Ford include lidar on its ICE vehicles if its so important a safety feature above the neural net and vision used by both humans and FSD to operate a car?

Is Ford producing a cyber taxi ?

Those seven words from Pinchai were not enough to assuage any doubt as to Tesla’s preeminence — for me anyway.

[Telecaster]

There are roughly 40,000 vehicular deaths a year in the US. This is a low bar for an AEV to clear. Of course we are not going to reduce deaths to zero, but if we can cut that number by 90% why wouldn't we do it? An AI driver will never drive distracted, tired, impaired or recklessly. An AI driver will never have road rage and cause an accident. What insurance company wouldn't insure an AEV if it was 90% less likely to be involved in a fatal accident compared to insuring a human driver? Cost of insuring will fall to the owner of the AEV, but cost of insuring each vehicle will be less than the cost of insuring a human driver and easily covered by revenue generated by a robotaxi.

Specifically, a Tesla operated with autopilot is 6X LESS likely to get in an accident compared to the US average.

https://www.tesla.com/VehicleSafetyReport

Actually it is a pretty high bar.  At least starting from where we are now.    The average human driver goes about 670K miles between collisions.
Meanwhile, FSD averages about 500 miles between disengagements.   Not all disengagements would lead to an accident of course, but even if 1 in 10 did, that is two orders of magnitude worse than a human.

But I think it is worse than that.  A disproportionately large percentage of accidents are caused by drunks and teenagers (or drunk teenagers).   Something like 35% of all fatalities involve alcohol. 
  I'm certain that regulators won't allow unsupervised FSD to be worse than an average, alert human and would probably require it to be better.     

I agree with what I believe is your main point, is that if FSD can be shown to be safer than the average alert human, then it would probably be mandated as safety equipment.    But there is a big gap from the current technology to that level and Tesla isn't the only company working on this.   

[GuitarStv]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

There are roughly 40,000 vehicular deaths a year in the US. This is a low bar for an AEV to clear. Of course we are not going to reduce deaths to zero, but if we can cut that number by 90% why wouldn't we do it? An AI driver will never drive distracted, tired, impaired or recklessly. An AI driver will never have road rage and cause an accident. What insurance company wouldn't insure an AEV if it was 90% less likely to be involved in a fatal accident compared to insuring a human driver? Cost of insuring will fall to the owner of the AEV, but cost of insuring each vehicle will be less than the cost of insuring a human driver and easily covered by revenue generated by a robotaxi.

Specifically, a Tesla operated with autopilot is 6X LESS likely to get in an accident compared to the US average.

https://www.tesla.com/VehicleSafetyReport

When death rates from driving fall in line with death rates from riding public transportation, flying on an airplane, or riding in a train, we will surrender control, just like we do in these other situations ,because we will deem the risk acceptable and better for society at large compared to the alternative of another 40,000 souls a year.

Those numbers from Tesla are kinda bullshit though, given the autopilot immediately cuts out nearly every time a vehicle is about to crash (hence why they say that the driver needs to monitor the vehicle constantly at all time that the 'self driving' is going on).  That's a clearly fudged slight of hand, not a reasonable estimation of the effectiveness of driverless operation.

Given that, it seems rather optimistic to use those numbers to determine insurance costs.

[ChpBstrd]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

Buddy, this is the country that once debated outlawing cigarettes, and then opted not to because it would hurt the stock market and eliminate jobs. We'll happily accept the sacrifice of a few thousand bicyclists and babies in strollers if it makes the number go up.

[Tigerpine]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

Buddy, this is the country that once debated outlawing cigarettes, and then opted not to because it would hurt the stock market and eliminate jobs. We'll happily accept the sacrifice of a few thousand bicyclists and babies in strollers if it makes the number go up.

Yes, but in all fairness, remember Prohibition?  That was actually implemented but didn't turn out so well in practice.

[ChpBstrd]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

Buddy, this is the country that once debated outlawing cigarettes, and then opted not to because it would hurt the stock market and eliminate jobs. We'll happily accept the sacrifice of a few thousand bicyclists and babies in strollers if it makes the number go up.

Yes, but in all fairness, remember Prohibition?  That was actually implemented but didn't turn out so well in practice.

It was arguably the last time the government banned an existing industry that was externalizing harms (aesbestos and lead paint/gasoline had easy alternatives). The bootleggers weren't as big a problem as the loss of tax revenue, and prohibition was repealed to pay for WW2. We've been trading lives for dollars and the political value of "more jobs" ever since.

[NorCal]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

Buddy, this is the country that once debated outlawing cigarettes, and then opted not to because it would hurt the stock market and eliminate jobs. We'll happily accept the sacrifice of a few thousand bicyclists and babies in strollers if it makes the number go up.

Yes, but in all fairness, remember Prohibition?  That was actually implemented but didn't turn out so well in practice.

It was arguably the last time the government banned an existing industry that was externalizing harms (aesbestos and lead paint/gasoline had easy alternatives). The bootleggers weren't as big a problem as the loss of tax revenue, and prohibition was repealed to pay for WW2. We've been trading lives for dollars and the political value of "more jobs" ever since.

This liability thing isn't even an issue with modern interpretations of consumer protections laws.

They'll make the passenger fully liable for all accidents and bury that in the terms & conditions that you have to accept to get a ride.  The mandatory arbitration clause will protect the company from class action lawsuits.

[bacchi]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

Buddy, this is the country that once debated outlawing cigarettes, and then opted not to because it would hurt the stock market and eliminate jobs. We'll happily accept the sacrifice of a few thousand bicyclists and babies in strollers if it makes the number go up.

Yes, but in all fairness, remember Prohibition?  That was actually implemented but didn't turn out so well in practice.

It was arguably the last time the government banned an existing industry that was externalizing harms (aesbestos and lead paint/gasoline had easy alternatives). The bootleggers weren't as big a problem as the loss of tax revenue, and prohibition was repealed to pay for WW2. We've been trading lives for dollars and the political value of "more jobs" ever since.

This liability thing isn't even an issue with modern interpretations of consumer protections laws.

They'll make the passenger fully liable for all accidents and bury that in the terms & conditions that you have to accept to get a ride.  The mandatory arbitration clause will protect the company from class action lawsuits.

Yeah but how long would that last? A few incidents and the company wouldn't get any fares.

"We can get an ArbitriCab, which might blame and then bankrupt us for its poor driving, or we can get a Waymo. What do you think?"

[NorCal]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

Buddy, this is the country that once debated outlawing cigarettes, and then opted not to because it would hurt the stock market and eliminate jobs. We'll happily accept the sacrifice of a few thousand bicyclists and babies in strollers if it makes the number go up.

Yes, but in all fairness, remember Prohibition?  That was actually implemented but didn't turn out so well in practice.

It was arguably the last time the government banned an existing industry that was externalizing harms (aesbestos and lead paint/gasoline had easy alternatives). The bootleggers weren't as big a problem as the loss of tax revenue, and prohibition was repealed to pay for WW2. We've been trading lives for dollars and the political value of "more jobs" ever since.

This liability thing isn't even an issue with modern interpretations of consumer protections laws.

They'll make the passenger fully liable for all accidents and bury that in the terms & conditions that you have to accept to get a ride.  The mandatory arbitration clause will protect the company from class action lawsuits.

Yeah but how long would that last? A few incidents and the company wouldn't get any fares.

"We can get an ArbitriCab, which might blame and then bankrupt us for its poor driving, or we can get a Waymo. What do you think?"

Just add on the optional insurance package for $12.99 per ride.  Just make sure it's not included in the price the customer sees before they request a ride. 

[ColoradoTribe]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

There are roughly 40,000 vehicular deaths a year in the US. This is a low bar for an AEV to clear. Of course we are not going to reduce deaths to zero, but if we can cut that number by 90% why wouldn't we do it? An AI driver will never drive distracted, tired, impaired or recklessly. An AI driver will never have road rage and cause an accident. What insurance company wouldn't insure an AEV if it was 90% less likely to be involved in a fatal accident compared to insuring a human driver? Cost of insuring will fall to the owner of the AEV, but cost of insuring each vehicle will be less than the cost of insuring a human driver and easily covered by revenue generated by a robotaxi.

Specifically, a Tesla operated with autopilot is 6X LESS likely to get in an accident compared to the US average.

https://www.tesla.com/VehicleSafetyReport

When death rates from driving fall in line with death rates from riding public transportation, flying on an airplane, or riding in a train, we will surrender control, just like we do in these other situations ,because we will deem the risk acceptable and better for society at large compared to the alternative of another 40,000 souls a year.

Those numbers from Tesla are kinda bullshit though, given the autopilot immediately cuts out nearly every time a vehicle is about to crash (hence why they say that the driver needs to monitor the vehicle constantly at all time that the 'self driving' is going on).  That's a clearly fudged slight of hand, not a reasonable estimation of the effectiveness of driverless operation.

Given that, it seems rather optimistic to use those numbers to determine insurance costs.

Fairly certain you are conflating Autopilot with FSD. Tesla numbers cited are for miles driven on autopilot, which is basically Tesla's driver assist system. Driver is in control but AI is running beneath the surface and intervenes to prevent accidents. FSD is when the car is driving but the driver needs to stay engaged/alert to take back control/intervene if necessary.

I driven two separate one month trials for FSD. The rate of improvement from the first trial to the second, with a few months in between, was astonishing. On the second version I was able to do entire drives, up to 40 minutes each way, with no interventions. Would I trust it enough to ride in the back seat, nope. However, during the second trial (late last year) I did find myself relaxing more and arriving at my destination less stressed than if I had driven it myself, which I find very telling.

[ColoradoTribe]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

Buddy, this is the country that once debated outlawing cigarettes, and then opted not to because it would hurt the stock market and eliminate jobs. We'll happily accept the sacrifice of a few thousand bicyclists and babies in strollers if it makes the number go up.

So unsafe, so many lives endangered by Tesla's disregard for safety.

https://www.youtube.com/watch?v=7Fhgx9zOTtY

[Runge]

What was the investment to R&D, build tooling, and produce the Cybertruck, and what is their current ROI?

What are the assumptions on Cybercab ROI?

The average cost of an Uber ride in the US is $2/mile. Let's assume a Cybercab undercuts this at $1.50/mile average. A Cybercab can operate 16 hours a day, leaving 8 hrs for recharging, maintenance and down time. Let's assume an average speed of 35 mph. So,  16hrs x 35 mph = 560 miles/day x 7days/week x 52 weeks = 203k miles/yr. Let's be super conservative and assume suboptimal utilization and each Cybercab does 100k billable miles per year. At $1.50/mile that's $150k in revenue per Cybercab per year. Fuel cost is roughly $0.15 per KwH. Cybercab gets 6 miles per KwH. So, annual fuel cost is $2,500 (assuming Tesla does provide their own energy via solar). Upkeep on a Cybercab is minimal. Cordless charging pads and robotic car cleaners take care of refueling and cleaning while charging. Maintenance cost is near zero. No scheduled maintenance in year one for a Tesla. Maybe some wiper fluid and one set of new tires. Let's put fueling, cleaning, registration fees and maintenance cost in year one at $5,000. Feel free to double that to $10k, it doesn't matter.

So, a Cybercab with a production cost of $15k rolls off the line and is accepting passengers the next day in Austin. In year 1 it generates $150k in revenue and costs $5k to operate, netting $145k for Tesla in year one. A nearly 10X ROI in the first year. The cab would pay for itself in under 2 months and produce enough profit in year 1 to build 6 additional Cybercabs.

Would like better sources, but claude.ai says that 30-40% of taxi miles driven are paid / with passengers and 60-70% are between fares. Some of those reasons wouldn't apply if there's no human, though many would:

Driving to and from initial pickup locations
Circling or waiting for fares
Returning to taxi stands
Repositioning between potential high-demand areas
Personal or maintenance-related driving

Obviously "returning to charging port" would be on there. Human taxi drivers seem to drive 45-70k, though depending on charge cycles, as well as peak and off-peak hours, it's fine to estimate 100k total miles being driven by an autonomous electric taxi.

Now assuming 50% of those are paid, because the machines are just better at algorithms than the existing humans + dispatch + software, you're looking at 100k miles, 50k in fares. Estimates for economy taxi rides in Austin, Tx are already around $1.50 / mile, so perhaps they undercut it at $1.25 / mile? And the only cost is $0.15 / mile electric charging. So over 100k miles, that's $15,000 in fuel cost, and $62,500 in revenue. (And as mentioned above, $5-10k in "other maintenance", hand-wavy napkin math!) If all of that checks out, it's still close to $40k / year profit per vehicle if all the assumptions are correct.

Here's a question though. Why are you questioning the economics of the Cybercab on an investor forum and asking others to do research that you could have done on your own before posting shade about the product?

Are the only people that can question if TSLA is a good investment, the people that are for investing? How about some contrary opinions so we can think about the things we haven't already? That goes both ways. If you're against it, why aren't you allowed to ask "why am I wrong, how could I be convinced this would be worth my investment dollars?"

I'm good with the above, while assuming some of your assumptions are on the low side and conceding my assumptions may be on the high side. Either way it is clear that Cybercab can be profitable in short order and undercut any competition. If Tesla bankrupts ride share services like Uber and Lyft and traditional taxi services, what does that do to utilization rates and billable miles per year per CC?

If the cost of using Tesla AEVs is so cheap as to undercut the cost of car ownership in urban and suburban locations what does that do to utilization rates and billable miles/year? If folks start forgoing car ownership and relying entirely on AEVs for ground transportation?

We need to be careful applying traditional metrics and assumptions to a disruptive technology. The ground is shifting under our feet.

I will give you the benefit of the doubt when it comes to the motivation behind your posts. You may very well be seeking objective understanding and contrarian opinions to debate and be open to an investment either way. Unfortunately, I have been on this thread since the beginning and some of the others who have been on for as long continue to move goal posts, dredge up the "concern of the day", refuse to concede when proven wrong, and whose single presence on this thread is to cheerlead against Tesla and amplify any FUD they happen across. Even though they've said time an again they would never own the stock.

I haven't commented on anything in YEARS, but I've lurked this whole time. I'd like to expand on a specific angle you mentioned regarding vehicle ownership.

You estimate that Tesla will be able to charge $1.5/mi to customers to ride.

I have two personal vehicles. One is my daily driver, which is an EV (non-Tesla, but comparable), the other is a ICE SUV. For the ICE vehicle, I've been tracking every expense and logging it in an app I've used for years. My current $/mi including all expenses and the capital cost to purchase the vehicle (used) is currently at $0.84/mi, and that's only after driving 43,000 miles over the past 3.5 years. My operating expenses to date have averaged out to $0.16/mi, so that $0.84/mi number will only go down the longer I keep it. If I'm able to keep it until I hit 100k miles, the average cost per mile will be $0.45/mi. At 200k if I can keep it that long? $0.31/mi. This is for a moderately efficient gas-powered SUV (26-30 mpg). Assumptions include no large deviations from maintenance costs and fuel costs. There are cheaper gas vehicles that I could have purchased to bring the captial costs down even further if I so desired.

For my EV. I purchased it back in October of last year, and I've put roughly 8k miles on it. Miles come mostly from my from commute but also 4-5 weekend road trips. It was not a cheap purchase at $45k, so that is working against me. I charge at home at $.13/kWh, that equates to $0.03/mi of fuel costs. Maintenance costs are hard for me to directly estimate, since I've paid $0 on maintenance to-date, but Google's search AI claims that the US DoE estimates a $0.07/mi average for maintenance of EVs. Going with that, then it's $0.10/mi for operating expenses. After 100k miles, this gives me a capital+operating expense average at $0.55/mi. If I can drive it until $200k, then that comes down to $0.33/mi. Roughly equivalent to my ICE vehicle (that cost half as much).

So doing some rough hand waving and (improperly) averaging both of those at 100k miles at $0.50/mi. My all-in expenses are 1/3 of the cost of using the Tesla's rideshare service using your $1.5/mi estimate. Why in the world would I ever make that switch? I'd be 3x'ing my transportation costs, plus eliminating my ability to drive long distance to visit family if I got rid of both vehicles. If I purchased even cheaper vehicles, then it would make even less sense.

Please explain how you see mass consumer adoption of robotaxi's without Tesla greatly reducing their margins. Otherwise, there's no financial incentive for people to switch fully over to rideshare, especially if a car you can own has the same self-driving capabilities as the robotaxi's.

[Kapyarn]

You estimate that Tesla will be able to charge $1.5/mi to customers to ride.

I have two personal vehicles. One is my daily driver, which is an EV (non-Tesla, but comparable), the other is a ICE SUV. For the ICE vehicle, I've been tracking every expense and logging it in an app I've used for years. My current $/mi including all expenses and the capital cost to purchase the vehicle (used) is currently at $0.84/mi, and that's only after driving 43,000 miles over the past 3.5 years. My operating expenses to date have averaged out to $0.16/mi, so that $0.84/mi number will only go down the longer I keep it. If I'm able to keep it until I hit 100k miles, the average cost per mile will be $0.45/mi. At 200k if I can keep it that long? $0.31/mi. This is for a moderately efficient gas-powered SUV (26-30 mpg). Assumptions include no large deviations from maintenance costs and fuel costs. There are cheaper gas vehicles that I could have purchased to bring the captial costs down even further if I so desired.

For my EV. I purchased it back in October of last year, and I've put roughly 8k miles on it. Miles come mostly from my from commute but also 4-5 weekend road trips. It was not a cheap purchase at $45k, so that is working against me. I charge at home at $.13/kWh, that equates to $0.03/mi of fuel costs. Maintenance costs are hard for me to directly estimate, since I've paid $0 on maintenance to-date, but Google's search AI claims that the US DoE estimates a $0.07/mi average for maintenance of EVs. Going with that, then it's $0.10/mi for operating expenses. After 100k miles, this gives me a capital+operating expense average at $0.55/mi. If I can drive it until $200k, then that comes down to $0.33/mi. Roughly equivalent to my ICE vehicle (that cost half as much).

So doing some rough hand waving and (improperly) averaging both of those at 100k miles at $0.50/mi. My all-in expenses are 1/3 of the cost of using the Tesla's rideshare service using your $1.5/mi estimate. Why in the world would I ever make that switch? I'd be 3x'ing my transportation costs, plus eliminating my ability to drive long distance to visit family if I got rid of both vehicles. If I purchased even cheaper vehicles, then it would make even less sense.

Please explain how you see mass consumer adoption of robotaxi's without Tesla greatly reducing their margins. Otherwise, there's no financial incentive for people to switch fully over to rideshare, especially if a car you can own has the same self-driving capabilities as the robotaxi's.

I am not 100% sure your numbers are correct.   I did not see any inclusion of car insurance for example.  For the average person, a $45,000 EV is going to have a car insurance full coverage rate of $2,000 to $3,000 a year.  If you drive 8000 miles a year gets you to $0.25 to $0.37 a mile just for insurance.

Also think about the opportunity cost of $45,000.  In the first year, you are missing $2,000 in interest from a 4.5% CD or T-bill.  That is another $0.25 a mile in opportunity cost.   Now you are at $0.50/mile to $0.63 a mile and we have just covered insurance and opportunity cost.

[neo von retorch]

Wait now we're going to assume the average consumer of transportation is going to do cost per mile analysis including opportunity cost? :)

Sorry, I disagree!

More realistically... people that use taxis and Taxi 2.0 (Uber / Lyft / Waymo) are... the same market that will use Taxi 3.0 (Waymo / Tesla / etc).

Convince me otherwise but Americans love their cars. Car purchase counts did decrease a little over COVID and with recent increasing "average" new car price (not to mention used car prices 2021-2022) but in general Americans love their cars and they want to "own" them (even if that means leasing them.)

The math has to be really stupidly good for them to give up their $600+ / month car payments and switch to taxis (if they haven't already because they are in the high density areas currently served by taxis.)

[Telecaster]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

I don't think this is really an issue.  There are deaths and injuries due to transporting people in general.   Airlines, bus companies, and taxis already kill people, so I don't see how eliminating a driver (or pilot) presents any particular new type of legal challenge.   

[Runge]

You estimate that Tesla will be able to charge $1.5/mi to customers to ride.

I have two personal vehicles. One is my daily driver, which is an EV (non-Tesla, but comparable), the other is a ICE SUV. For the ICE vehicle, I've been tracking every expense and logging it in an app I've used for years. My current $/mi including all expenses and the capital cost to purchase the vehicle (used) is currently at $0.84/mi, and that's only after driving 43,000 miles over the past 3.5 years. My operating expenses to date have averaged out to $0.16/mi, so that $0.84/mi number will only go down the longer I keep it. If I'm able to keep it until I hit 100k miles, the average cost per mile will be $0.45/mi. At 200k if I can keep it that long? $0.31/mi. This is for a moderately efficient gas-powered SUV (26-30 mpg). Assumptions include no large deviations from maintenance costs and fuel costs. There are cheaper gas vehicles that I could have purchased to bring the captial costs down even further if I so desired.

For my EV. I purchased it back in October of last year, and I've put roughly 8k miles on it. Miles come mostly from my from commute but also 4-5 weekend road trips. It was not a cheap purchase at $45k, so that is working against me. I charge at home at $.13/kWh, that equates to $0.03/mi of fuel costs. Maintenance costs are hard for me to directly estimate, since I've paid $0 on maintenance to-date, but Google's search AI claims that the US DoE estimates a $0.07/mi average for maintenance of EVs. Going with that, then it's $0.10/mi for operating expenses. After 100k miles, this gives me a capital+operating expense average at $0.55/mi. If I can drive it until $200k, then that comes down to $0.33/mi. Roughly equivalent to my ICE vehicle (that cost half as much).

So doing some rough hand waving and (improperly) averaging both of those at 100k miles at $0.50/mi. My all-in expenses are 1/3 of the cost of using the Tesla's rideshare service using your $1.5/mi estimate. Why in the world would I ever make that switch? I'd be 3x'ing my transportation costs, plus eliminating my ability to drive long distance to visit family if I got rid of both vehicles. If I purchased even cheaper vehicles, then it would make even less sense.

Please explain how you see mass consumer adoption of robotaxi's without Tesla greatly reducing their margins. Otherwise, there's no financial incentive for people to switch fully over to rideshare, especially if a car you can own has the same self-driving capabilities as the robotaxi's.

I am not 100% sure your numbers are correct.   I did not see any inclusion of car insurance for example.  For the average person, a $45,000 EV is going to have a car insurance full coverage rate of $2,000 to $3,000 a year.  If you drive 8000 miles a year gets you to $0.25 to $0.37 a mile just for insurance.

Also think about the opportunity cost of $45,000.  In the first year, you are missing $2,000 in interest from a 4.5% CD or T-bill.  That is another $0.25 a mile in opportunity cost.   Now you are at $0.50/mile to $0.63 a mile and we have just covered insurance and opportunity cost.

You are correct that I didn't include insurance (my laziness). I looked up the numbers. For both vehicles together, I'm paying $2800/year. The EV is 1500/year with full coverage. I don't drive 8k/year as I've put 8k on it within just the past 6 months. So if I continue driving the same amount, then I'm at 16k miles per year. $1500/16000mi is $0.09/mi. So change my previous estimate from $0.55/mi to $0.65/mi and I'm still only spending 43% of the cost if I fully switched over to robotaxi's at $1.5/mi.

If I was more cost conscious, I could purchase a much cheaper EV at $15k (such as a used Bolt) and that would make even more in favor of my owning the vehicle myself. A 15k EV with $0.10/mi fuel+maintenance and $0.10/mi insurance comes out to $0.28/mi.

Edit: Or better yet, if I was MMM cost-conscious instead of my consumer sucker self, then I would move closer to work and ride my bike. But that's a different conversation.

[dividendman]

Wait now we're going to assume the average consumer of transportation is going to do cost per mile analysis including opportunity cost? :)

Sorry, I disagree!

More realistically... people that use taxis and Taxi 2.0 (Uber / Lyft / Waymo) are... the same market that will use Taxi 3.0 (Waymo / Tesla / etc).

Convince me otherwise but Americans love their cars. Car purchase counts did decrease a little over COVID and with recent increasing "average" new car price (not to mention used car prices 2021-2022) but in general Americans love their cars and they want to "own" them (even if that means leasing them.)

The math has to be really stupidly good for them to give up their $600+ / month car payments and switch to taxis (if they haven't already because they are in the high density areas currently served by taxis.)

Ain't nobody drivin' 'round in a gubment tracked taxi with no god dang driver! Ya'll crazy.

[neo von retorch]

Ain't nobody drivin' 'round in a gubment tracked taxi with no god dang driver! Ya'll crazy.

Well, can't do much better in your own car...

[ColoradoTribe]

You estimate that Tesla will be able to charge $1.5/mi to customers to ride.

I have two personal vehicles. One is my daily driver, which is an EV (non-Tesla, but comparable), the other is a ICE SUV. For the ICE vehicle, I've been tracking every expense and logging it in an app I've used for years. My current $/mi including all expenses and the capital cost to purchase the vehicle (used) is currently at $0.84/mi, and that's only after driving 43,000 miles over the past 3.5 years. My operating expenses to date have averaged out to $0.16/mi, so that $0.84/mi number will only go down the longer I keep it. If I'm able to keep it until I hit 100k miles, the average cost per mile will be $0.45/mi. At 200k if I can keep it that long? $0.31/mi. This is for a moderately efficient gas-powered SUV (26-30 mpg). Assumptions include no large deviations from maintenance costs and fuel costs. There are cheaper gas vehicles that I could have purchased to bring the captial costs down even further if I so desired.

For my EV. I purchased it back in October of last year, and I've put roughly 8k miles on it. Miles come mostly from my from commute but also 4-5 weekend road trips. It was not a cheap purchase at $45k, so that is working against me. I charge at home at $.13/kWh, that equates to $0.03/mi of fuel costs. Maintenance costs are hard for me to directly estimate, since I've paid $0 on maintenance to-date, but Google's search AI claims that the US DoE estimates a $0.07/mi average for maintenance of EVs. Going with that, then it's $0.10/mi for operating expenses. After 100k miles, this gives me a capital+operating expense average at $0.55/mi. If I can drive it until $200k, then that comes down to $0.33/mi. Roughly equivalent to my ICE vehicle (that cost half as much).

So doing some rough hand waving and (improperly) averaging both of those at 100k miles at $0.50/mi. My all-in expenses are 1/3 of the cost of using the Tesla's rideshare service using your $1.5/mi estimate. Why in the world would I ever make that switch? I'd be 3x'ing my transportation costs, plus eliminating my ability to drive long distance to visit family if I got rid of both vehicles. If I purchased even cheaper vehicles, then it would make even less sense.

Please explain how you see mass consumer adoption of robotaxi's without Tesla greatly reducing their margins. Otherwise, there's no financial incentive for people to switch fully over to rideshare, especially if a car you can own has the same self-driving capabilities as the robotaxi's.

I am not 100% sure your numbers are correct.   I did not see any inclusion of car insurance for example.  For the average person, a $45,000 EV is going to have a car insurance full coverage rate of $2,000 to $3,000 a year.  If you drive 8000 miles a year gets you to $0.25 to $0.37 a mile just for insurance.

Also think about the opportunity cost of $45,000.  In the first year, you are missing $2,000 in interest from a 4.5% CD or T-bill.  That is another $0.25 a mile in opportunity cost.   Now you are at $0.50/mile to $0.63 a mile and we have just covered insurance and opportunity cost.

I also don't see vehicle depreciation. You are also extrapolating maintenance costs for the life of the vehicle using only routine maintenance to date. These costs are lumpy and unexpected. A broken windshield, the fender bender that's not worth an insurance claim, transmission flush, new muffler, timing belt, etc. Do you wash your car? Even at home that costs something. Ever drive a toll road? Is the car financed? Most people will have interest in addition to the opportunity cost of investing the money spent on the car's purchase price.

In big cities there is the cost of even storing/parking your car.

How much value do we put on our time. Time not spent shopping insurance, washing and vacuuming the car, oil changes, registration, emissions testing, vehicle inspections, etc.

No one can steal, break into or vandalize something you don't own. Ride hailing an AEV instead of owning a car clears up a lot of physical space and mental space, which is a tremendous value in my opinion.

My guess, a lot of folks will opt out of car ownership when a proven, reliable AEV service is available and costs around $1.00/mile. Tesla could operate a profitable Cybercab for that amount based on the economics laid out above.

[ColoradoTribe]

Wait now we're going to assume the average consumer of transportation is going to do cost per mile analysis including opportunity cost? :)

Sorry, I disagree!

More realistically... people that use taxis and Taxi 2.0 (Uber / Lyft / Waymo) are... the same market that will use Taxi 3.0 (Waymo / Tesla / etc).

Convince me otherwise but Americans love their cars. Car purchase counts did decrease a little over COVID and with recent increasing "average" new car price (not to mention used car prices 2021-2022) but in general Americans love their cars and they want to "own" them (even if that means leasing them.)

The math has to be really stupidly good for them to give up their $600+ / month car payments and switch to taxis (if they haven't already because they are in the high density areas currently served by taxis.)

Rates of car ownership are already going down with younger generations before the option of an AEV is available. Gas on a fire. Car ownership today is not the same "joy of the open road" experience older generations associate with car ownership. Nor is it the status symbol with the younger crowd that it is with older generations.

[neo von retorch]

Rates of car ownership are already going down with younger generations before the option of an AEV is available. Gas on a fire. Car ownership today is not the same "joy of the open road" experience older generations associate with car ownership. Nor is it the status symbol with the younger crowd that it is with older generations.

https://www.forbes.com/advisor/car-insurance/car-ownership-statistics/

The number of registered vehicles in the United States increased by 3.5% between 2018 and 2022, from 269,417,884 registered vehicles to 278,870,463 registered vehicles, indicating an upward trend in car ownership.

Car ownership in America is on the rise. Only 8.3% of households did not have a vehicle in 2022, a 4.6% decrease from 2018, when 8.7% of households did not have a vehicle.

https://www.fool.com/money/research/car-ownership-statistics/

Rising car ownership: 91.7% of U.S. households had at least one vehicle in 2022, up from 90.9% in 2015.

https://www.consumeraffairs.com/automotive/car-ownership-statistics.html

In 2022, sales of new light vehicles dropped to 13.9 million units, down from 14.6 million units in 2021.
Sales rebounded in 2023, with 15.5 million light vehicles sold in the U.S., including 3.12 million passenger cars and 12.4 million light trucks.

https://www.autoinsurance.com/research/car-ownership-statistics/

Based on the latest United States Census data, 92 percent of American households own at least one car.

I'm not seeing enough evidence to agree with your premise here.

GENERATIONAL TRENDS IN VEHICLE OWNERSHIP AND USE: ARE MILLENNIALS ANY DIFFERENT? (PDF, 2019)

We find that although a simple comparison of average ownership and use would suggest a difference, once one controls for confounding variables there is no evidence of a difference. While we find that Millennials are altering life-choices that affect vehicle ownership, the net effect of these endogenous choices is to reduce vehicle ownership by less than one percent. We can statistically rule out effects larger than two percent.

Closest I can get to support... https://www.foley.com/insights/publications/2023/05/data-deep-dive-gen-z-less-interested-owning-car/

While it is difficult to predict with certainty how car sales will be impacted by Gen Z’s lack of interest in cars and driving, a deeper dive into Gen Z’s current lifestyle choices and feelings towards driving shows their ambivalence towards cars may be short-lived. For example, as Gen Z grows up, moves away from city centers and college campuses, and has children, their need for day-to-day private transportation will likely increase. Additionally, while ride-sharing services are popular for this age group, they may not be as cost-effective in the long term as owning a car, especially for individuals who travel frequently and need to commute long distances. Moreover, as electric and hybrid vehicles become more affordable and accessible, they may become an appealing option to the environmentally conscious Gen Z. Further, developments in autonomous driving technology and enhanced safety features may alleviate some of the anxiety associated with driving that Gen Z currently reports.

A trip back to when Millennials began reaching driving age is instructive. A decade ago it appears that Millennials were similarly disinterested in owning a car in comparison to earlier generations, like the boomers. For example, the number of young Americans driving in 2014 significantly declined from earlier generations and the prevalent narrative was that Millennials rejected cars. However, as Millennials grew up, it turned out that their attitudes toward car culture were not so different from the boomer generation. It was not but a few years later, as Millennials aged, began to establish their careers, and were able to finally afford cars with their raising incomes that a larger percentage began purchasing cars just like their parents had

[Runge]

You estimate that Tesla will be able to charge $1.5/mi to customers to ride.

I have two personal vehicles. One is my daily driver, which is an EV (non-Tesla, but comparable), the other is a ICE SUV. For the ICE vehicle, I've been tracking every expense and logging it in an app I've used for years. My current $/mi including all expenses and the capital cost to purchase the vehicle (used) is currently at $0.84/mi, and that's only after driving 43,000 miles over the past 3.5 years. My operating expenses to date have averaged out to $0.16/mi, so that $0.84/mi number will only go down the longer I keep it. If I'm able to keep it until I hit 100k miles, the average cost per mile will be $0.45/mi. At 200k if I can keep it that long? $0.31/mi. This is for a moderately efficient gas-powered SUV (26-30 mpg). Assumptions include no large deviations from maintenance costs and fuel costs. There are cheaper gas vehicles that I could have purchased to bring the captial costs down even further if I so desired.

For my EV. I purchased it back in October of last year, and I've put roughly 8k miles on it. Miles come mostly from my from commute but also 4-5 weekend road trips. It was not a cheap purchase at $45k, so that is working against me. I charge at home at $.13/kWh, that equates to $0.03/mi of fuel costs. Maintenance costs are hard for me to directly estimate, since I've paid $0 on maintenance to-date, but Google's search AI claims that the US DoE estimates a $0.07/mi average for maintenance of EVs. Going with that, then it's $0.10/mi for operating expenses. After 100k miles, this gives me a capital+operating expense average at $0.55/mi. If I can drive it until $200k, then that comes down to $0.33/mi. Roughly equivalent to my ICE vehicle (that cost half as much).

So doing some rough hand waving and (improperly) averaging both of those at 100k miles at $0.50/mi. My all-in expenses are 1/3 of the cost of using the Tesla's rideshare service using your $1.5/mi estimate. Why in the world would I ever make that switch? I'd be 3x'ing my transportation costs, plus eliminating my ability to drive long distance to visit family if I got rid of both vehicles. If I purchased even cheaper vehicles, then it would make even less sense.

Please explain how you see mass consumer adoption of robotaxi's without Tesla greatly reducing their margins. Otherwise, there's no financial incentive for people to switch fully over to rideshare, especially if a car you can own has the same self-driving capabilities as the robotaxi's.

I am not 100% sure your numbers are correct.   I did not see any inclusion of car insurance for example.  For the average person, a $45,000 EV is going to have a car insurance full coverage rate of $2,000 to $3,000 a year.  If you drive 8000 miles a year gets you to $0.25 to $0.37 a mile just for insurance.

Also think about the opportunity cost of $45,000.  In the first year, you are missing $2,000 in interest from a 4.5% CD or T-bill.  That is another $0.25 a mile in opportunity cost.   Now you are at $0.50/mile to $0.63 a mile and we have just covered insurance and opportunity cost.

I also don't see vehicle depreciation. You are also extrapolating maintenance costs for the life of the vehicle using only routine maintenance to date. These costs are lumpy and unexpected. A broken windshield, the fender bender that's not worth an insurance claim, transmission flush, new muffler, timing belt, etc. Do you wash your car? Even at home that costs something. Ever drive a toll road? Is the car financed? Most people will have interest in addition to the opportunity cost of investing the money spent on the car's purchase price.

In big cities there is the cost of even storing/parking your car.

How much value do we put on our time. Time not spent shopping insurance, washing and vacuuming the car, oil changes, registration, emissions testing, vehicle inspections, etc.

No one can steal, break into or vandalize something you don't own. Ride hailing an AEV instead of owning a car clears up a lot of physical space and mental space, which is a tremendous value in my opinion.

My guess, a lot of folks will opt out of car ownership when a proven, reliable AEV service is available and costs around $1.00/mile. Tesla could operate a profitable Cybercab for that amount based on the economics laid out above.

You gives a lot of hand waving and speculation, and you just moved the goalpost to $1.00/mi. People TODAY don't like taking the local managed/tolled lanes that let them bypass heavy traffic over 10 miles for only $1.

My numbers include tolls as part of my fuel costs, although they're minimal since I rarely use tolls. Tolls would be passed onto the ride hailer anyways like they do now. The point is moot.

Cost of storing a car is baked into our existing housing since every home is required to have one or two parking spots. That isn't materially changing within the next 10 years.

Vehicle depreciation in my numbers is assumed to go to zero. If I sell the vehicle, then the actual depreciation is less than my assumption, making the comparison between owning vs renting even more in favor of owning.

As for time. I do wash it, doesn't take long or cost much to run through it every other month or so. Shopping for insurance is maybe an hour per year, if that. There's no oil changes, transmission flushes, timing belts, mufflers, etc on EVs. What's your point? For maintenance I drop it off at the dealer and someone picks me up so I go to work while they work on the vehicle. My state no longer requires inspections, and EVs don't emit anything. None of those things are that big of a hurdle, especially with EVs. If the alternative mode is 2x the cost of owning my vehicle, then I'll gladly accept owning my vehicle.

Speaking for myself...again if I was really cost conscious, then I would move closer to work, or get a job closer to where I live and ride a bike. Negating car ownership/usage at all. And if I did use a ride hail, I would not be logging 16k miles per year.

[Kapyarn]

16k miles a year is different.  At that level, owning a vehicle makes a lot more sense.

Someone who does 5k miles a year though, I could see the insurance and vehicle costs driving a large portion of the ownership value.

$5,000 to $7,500 a year for robo taxi and never having to pay insurance, fuel, parking, or even wash it?  Could appeal to a lot of people.

[NorCal]

As much as I believe robotaxi whatever is a nonsense business designed to pump the stock while creating little of value, there is a new consideration that might bring some real value to the business.

The new tariff structure (assuming it stays) is about to make car ownership out of reach for a lot of Americans.  There might be a business not because people want it, but because they can no longer afford car ownership.

[ColoradoTribe]

Rates of car ownership are already going down with younger generations before the option of an AEV is available. Gas on a fire. Car ownership today is not the same "joy of the open road" experience older generations associate with car ownership. Nor is it the status symbol with the younger crowd that it is with older generations.

https://www.forbes.com/advisor/car-insurance/car-ownership-statistics/

The number of registered vehicles in the United States increased by 3.5% between 2018 and 2022, from 269,417,884 registered vehicles to 278,870,463 registered vehicles, indicating an upward trend in car ownership.

Car ownership in America is on the rise. Only 8.3% of households did not have a vehicle in 2022, a 4.6% decrease from 2018, when 8.7% of households did not have a vehicle.

https://www.fool.com/money/research/car-ownership-statistics/

Rising car ownership: 91.7% of U.S. households had at least one vehicle in 2022, up from 90.9% in 2015.

https://www.consumeraffairs.com/automotive/car-ownership-statistics.html

In 2022, sales of new light vehicles dropped to 13.9 million units, down from 14.6 million units in 2021.
Sales rebounded in 2023, with 15.5 million light vehicles sold in the U.S., including 3.12 million passenger cars and 12.4 million light trucks.

https://www.autoinsurance.com/research/car-ownership-statistics/

Based on the latest United States Census data, 92 percent of American households own at least one car.

I'm not seeing enough evidence to agree with your premise here.

GENERATIONAL TRENDS IN VEHICLE OWNERSHIP AND USE: ARE MILLENNIALS ANY DIFFERENT? (PDF, 2019)

We find that although a simple comparison of average ownership and use would suggest a difference, once one controls for confounding variables there is no evidence of a difference. While we find that Millennials are altering life-choices that affect vehicle ownership, the net effect of these endogenous choices is to reduce vehicle ownership by less than one percent. We can statistically rule out effects larger than two percent.

Closest I can get to support... https://www.foley.com/insights/publications/2023/05/data-deep-dive-gen-z-less-interested-owning-car/

While it is difficult to predict with certainty how car sales will be impacted by Gen Z’s lack of interest in cars and driving, a deeper dive into Gen Z’s current lifestyle choices and feelings towards driving shows their ambivalence towards cars may be short-lived. For example, as Gen Z grows up, moves away from city centers and college campuses, and has children, their need for day-to-day private transportation will likely increase. Additionally, while ride-sharing services are popular for this age group, they may not be as cost-effective in the long term as owning a car, especially for individuals who travel frequently and need to commute long distances. Moreover, as electric and hybrid vehicles become more affordable and accessible, they may become an appealing option to the environmentally conscious Gen Z. Further, developments in autonomous driving technology and enhanced safety features may alleviate some of the anxiety associated with driving that Gen Z currently reports.

A trip back to when Millennials began reaching driving age is instructive. A decade ago it appears that Millennials were similarly disinterested in owning a car in comparison to earlier generations, like the boomers. For example, the number of young Americans driving in 2014 significantly declined from earlier generations and the prevalent narrative was that Millennials rejected cars. However, as Millennials grew up, it turned out that their attitudes toward car culture were not so different from the boomer generation. It was not but a few years later, as Millennials aged, began to establish their careers, and were able to finally afford cars with their raising incomes that a larger percentage began purchasing cars just like their parents had

My takeaway from your post. Car ownership is expensive and millennials delayed car ownership or were trying to avoid it until life mandated the needed one. Either way, what if the next generation uses an AEV service while waiting to afford their own car and realizes they're getting along just fine without their own car and saving a lot of money in the process?

Again, this is the nature of disruption. People struggle to see another reality than the one they grew up, b ut if the next generation grows up with on demand ride hailing from an AEV this will be the only reality the know and they will view car ownership as wasteful and antiquated.

[Runge]

snip

My takeaway from your post. Car ownership is expensive and millennials delayed car ownership or were trying to avoid it until life mandated the needed one. Either way, what if the next generation uses an AEV service while waiting to afford their own car and realizes they're getting along just fine without their own car and saving a lot of money in the process?

Again, this is the nature of disruption. People struggle to see another reality than the one they grew up, b ut if the next generation grows up with on demand ride hailing from an AEV this will be the only reality the know and they will view car ownership as wasteful and antiquated.

You keep saying that ride hailing will be cheaper than owning, but have yet to provide any evidence to support your claims. Yes car ownership is expensive, but all evidence to-date is that ride hailing is MORE expensive than owning. And the estimates you gave (no idea where you got them), indicate that ride hailing will be 2-3x more expensive than owning, if I use my numbers.

[neo von retorch]

My takeaway from your post. Car ownership is expensive and millennials delayed car ownership or were trying to avoid it until life mandated the needed one. Either way, what if the next generation uses an AEV service while waiting to afford their own car and realizes they're getting along just fine without their own car and saving a lot of money in the process?

Again, this is the nature of disruption. People struggle to see another reality than the one they grew up, b ut if the next generation grows up with on demand ride hailing from an AEV this will be the only reality the know and they will view car ownership as wasteful and antiquated.

I think disruption is a bit hand-wavy.

Especially when you remember that the U.S. is huge and very spread out, with lots of the population living in low density areas.

At best, you'll get some percentage of the next generation that would have otherwise bought a car choosing to, instead, switch to an ride-hailing service (not unlike the small percentage that felt that Uber / Lyft was worth using when taxi cabs were not.) If you're going for real Total Addressable Market (TAM) analysis, then you've got to ask some questions:

- what's the current ride-hailing market?
- how much of that market will readily convert to autonomous vehicles, and at what price point?
- what's the current car-ownership market?
- in what locations is the density of population and destinations great enough to shift consumers from car-ownership to ride-hailing... but is not currently served by ride-hailing services?
- how many total consumers are in that market, and how many can be convinced to park with their cars?
- in less dense areas, how much additional unpaid driving will autonomous vehicles have to travel between fares, affecting margins?

Some are easy questions, and some are very hard, and are at the very core of total potential revenue and profits of the theoretical future.

Beyond that, of course, are huge questions about timing.

- how quickly will software catch up to the 2016 forecast of full self-driving level required for truly autonomous ride-hailing services?
- how quickly will manufacturing ramp up? (will it follow the Model 3, the Model Y, the Cybertruck? what unforeseen hiccups will happen along the way?)
- how quickly will competitors do the same things, within some margin of "competitiveness", and how much of the TAM will they take?
- how quickly will regulation hurdles be overcome?
- how quickly will infrastructure for "self-charging" cars be built and distributed sufficiently?
- what is the margin for Uber / Lyft / etc. and how much can they adapt to changing market conditions? (what happened to taxi cab companies when Uber and Lyft became popular?)

[Telecaster]

More realistically... people that use taxis and Taxi 2.0 (Uber / Lyft / Waymo) are... the same market that will use Taxi 3.0 (Waymo / Tesla / etc).

Totally agree, with the slight caveat that if Taxi 3.0 was cheaper and more convenient than Taxi 2.0 then it makes sense more people would use it. 

 

I also don't see vehicle depreciation. You are also extrapolating maintenance costs for the life of the vehicle using only routine maintenance to date. These costs are lumpy and unexpected. A broken windshield, the fender bender that's not worth an insurance claim, transmission flush, new muffler, timing belt, etc. Do you wash your car? Even at home that costs something. Ever drive a toll road? Is the car financed? Most people will have interest in addition to the opportunity cost of investing the money spent on the car's purchase price.

In big cities there is the cost of even storing/parking your car.

How much value do we put on our time. Time not spent shopping insurance, washing and vacuuming the car, oil changes, registration, emissions testing, vehicle inspections, etc.

No one can steal, break into or vandalize something you don't own. Ride hailing an AEV instead of owning a car clears up a lot of physical space and mental space, which is a tremendous value in my opinion.

My guess, a lot of folks will opt out of car ownership when a proven, reliable AEV service is available and costs around $1.00/mile. Tesla could operate a profitable Cybercab for that amount based on the economics laid out above.

The problem is that all those same costs exist for the rideshare company too.   Only you do the light maintenance work for free, but the rideshare company needs to pay somebody.  You park in your driveway for free, but the rideshare company needs to lease a parking lot, and so on.   And on top of that the rideshare company needs to make a profit.   

I easily can see an autonomous rideshare company being cheaper than one with human drivers, but I don't see how it works out to be meaningfully cheaper than owning the equivalent vehicle.   

[GuitarStv]

Even the most optimistic folks agree that there will be deaths and injury due to automated driving.  Are the costs associated with those deaths accounted for in the profits?  How does one insure an automated vehicle?

I don't think this is really an issue.  There are deaths and injuries due to transporting people in general.   Airlines, bus companies, and taxis already kill people, so I don't see how eliminating a driver (or pilot) presents any particular new type of legal challenge.

Insurance is calculated based upon risk.  We don't have a risk profile for automated vehicle travel yet.  I guess it depends on whether you believe Tesla's fudged numbers or the results of the NHTSA investigations.  In the former, insurance will be cheaper for autonomous vehicles than people.  In the latter, insurance will be more expensive than for a human driver (if it is provided at all).  Waymo has kinda sidestepped this problem by having a staff constantly monitoring the 'self-driving' vehicles and intervening on a regular basis.  My understanding was that Tesla was planning on having the vehicles be fully autonomous, so would necessarily be treated differently.

[Posthumane]

Apart from the cost aspects, there are a number of reasons people own cars over using taxis or ride hailing services, or even public transportation. A lot of the time when I take my car somewhere I take a bunch of stuff with me that I will want to have available to me on my trip or at my destination, but not stuff that I want to carry around with me all the time. You can't do that with a taxi service. You also can't use a taxi service to go tow a trailer or truck camper, and many people I know own excessively large/expensive vehicles specifically for those reasons. Basically there are enough reasons for many people to own vehicles that only a small number of households in North American would be go entirely car free even if a much cheaper ride hailing service was available. And once you own a vehicle and are already paying the insurance/parking costs on it, the marginal cost of using that vehicle is always going to be lower than a for-profit ride hailing service.

To sum up: the conclusion that people will switch from car ownership to robotaxi use if robotaxis are significantly cheaper that Uber et al. is based on a false premise that people own cars now only because they are cheaper than using taxis.

[ColoradoTribe]

More realistically... people that use taxis and Taxi 2.0 (Uber / Lyft / Waymo) are... the same market that will use Taxi 3.0 (Waymo / Tesla / etc).

Totally agree, with the slight caveat that if Taxi 3.0 was cheaper and more convenient than Taxi 2.0 then it makes sense more people would use it. 

 

I also don't see vehicle depreciation. You are also extrapolating maintenance costs for the life of the vehicle using only routine maintenance to date. These costs are lumpy and unexpected. A broken windshield, the fender bender that's not worth an insurance claim, transmission flush, new muffler, timing belt, etc. Do you wash your car? Even at home that costs something. Ever drive a toll road? Is the car financed? Most people will have interest in addition to the opportunity cost of investing the money spent on the car's purchase price.

In big cities there is the cost of even storing/parking your car.

How much value do we put on our time. Time not spent shopping insurance, washing and vacuuming the car, oil changes, registration, emissions testing, vehicle inspections, etc.

No one can steal, break into or vandalize something you don't own. Ride hailing an AEV instead of owning a car clears up a lot of physical space and mental space, which is a tremendous value in my opinion.

My guess, a lot of folks will opt out of car ownership when a proven, reliable AEV service is available and costs around $1.00/mile. Tesla could operate a profitable Cybercab for that amount based on the economics laid out above.

The problem is that all those same costs exist for the rideshare company too.   Only you do the light maintenance work for free, but the rideshare company needs to pay somebody.  You park in your driveway for free, but the rideshare company needs to lease a parking lot, and so on.   And on top of that the rideshare company needs to make a profit.   

I easily can see an autonomous rideshare company being cheaper than one with human drivers, but I don't see how it works out to be meaningfully cheaper than owning the equivalent vehicle.

Really, I don't know. Average new car price in the US is approaching $40k. Average that out over 10 years of ownership. That's $4k/yr. Then add the opportunity cost of buying a depreciating asset versus investing $40k. Then add annual tax, title, registration, insurance, maintenance, and fuel costs. Owning a car costs around 8-10k/yr. As shown above, a Cybercab can profitably be operated for $1/mile.

How many Americans drive less than 8-10K miles/yr?

BTW, Musk is quoted as saying the cost to operate a CC is 20 cents/mile. Even if you double that to 40 cents, it would mean the cost of car ownership would be more expensive than using AEVs for transport unless you were driving over 20,000 miles a year (or lived in a very rural area, which is a small minority of the pop). Owning your own vehicle could become a luxury only justifiable to the wealthy in suburban and urban markets.

[Telecaster]

I think disruption is a bit hand-wavy.

Especially when you remember that the U.S. is huge and very spread out, with lots of the population living in low density areas.

"Disruption" is a lot hand wavy.   Ridesharing is an existing business.   This would simply be automating away the human driver.   That's an incremental change in an industry that has existed before there were even automobiles.   

A ridesharing business only makes sense in fairly dense areas.    Removing the human driver would presumably drop costs, and therefore make it more economical in less dense areas.   But there is still a limit.   

And not to put too fine a point on all this, the technology required for Level 5 autonomy does not exist and there is no timeline for when it may exist.   There seems to be a lot chicken counting going on.