Tesla Network / Robotaxi Business Model

[Mengy]

I can't imagine more than a few percent of people would opt to make their Tesla a robotaxi. Seems like Elon is assuming it would be almost everyone.

I agree. I know I would not, and every Tesla owner I've discussed this with aren't planning to either.

This is why I frown whenever I see Super Tesla Bulls talk about Level 5 FSD equating to instant revenues and a soaring stock price. My opinion is it won't play out like that. While solved FSD would be a huge breakthrough, and yes it would eventually lead to robotaxi fleets earning lots of revenue for Tesla, that will not happen instantly nor quickly. The fleet will need to be deployed (because most Tesla car owners probably won't be sharing their vehicle with the public), a new robotaxi vehicle will need to be developed / production lines built / manufactured, robotaxi service and parking centers need to be built, infrastructure needs to be deployed, regulations need to be passed, etc.

I also don't see tons of people rushing to spend $15,000+ for Level 5 FSD either. Yes some people will, but the majority of Tesla owners likely won't be paying that price tag when they can simply drive themselves and use the free Autopilot. $15,000+ for software is not a bill the majority of people will pay in this world and economy.

Solved autonomy will be a slow, gradual adder to TSLA and Tesla financials. Sizeable in time yes, but in a LONG time. IMHO.

 

[tivoboy]

I agree. I know I would not, and every Tesla owner I've discussed this with aren't planning to either.

This is why I frown whenever I see Super Tesla Bulls talk about Level 5 FSD equating to instant revenues and a soaring stock price. My opinion is it won't play out like that. While solved FSD would be a huge breakthrough, and yes it would eventually lead to robotaxi fleets earning lots of revenue for Tesla, that will not happen instantly nor quickly. The fleet will need to be deployed (because most Tesla car owners probably won't be sharing their vehicle with the public), a new robotaxi vehicle will need to be developed / production lines built / manufactured, robotaxi service and parking centers need to be built, infrastructure needs to be deployed, regulations need to be passed, etc.

I also don't see tons of people rushing to spend $15,000+ for Level 5 FSD either. Yes some people will, but the majority of Tesla owners likely won't be paying that price tag when they can simply drive themselves and use the free Autopilot. $15,000+ for software is not a bill the majority of people will pay in this world and economy.

Solved autonomy will be a slow, gradual adder to TSLA and Tesla financials. Sizeable in time yes, but in a LONG time. IMHO.

I think once, IF Tesla can ever get FSD to Level 5 autonomy, (or really just level 4 for that matter), the unlock price will no longer be 15K, probably 50-100% higher IMHO.

 

[ItsNotAboutTheMoney]

I agree. I know I would not, and every Tesla owner I've discussed this with aren't planning to either.

This is why I frown whenever I see Super Tesla Bulls talk about Level 5 FSD equating to instant revenues and a soaring stock price. My opinion is it won't play out like that. While solved FSD would be a huge breakthrough, and yes it would eventually lead to robotaxi fleets earning lots of revenue for Tesla, that will not happen instantly nor quickly. The fleet will need to be deployed (because most Tesla car owners probably won't be sharing their vehicle with the public), a new robotaxi vehicle will need to be developed / production lines built / manufactured, robotaxi service and parking centers need to be built, infrastructure needs to be deployed, regulations need to be passed, etc.

There's about 1.7M rideshare drivers in the USA and 230 taxi drivers.
Fremont is at 550k/year, Austin is approaching 250k/year.

If Tesla had Level 5, the scaling limitation would largely be service. If you can build lots of new cars, no rush to get existing vehicles fixed.

I also don't see tons of people rushing to spend $15,000+ for Level 5 FSD either. Yes some people will, but the majority of Tesla owners likely won't be paying that price tag when they can simply drive themselves and use the free Autopilot. $15,000+ for software is not a bill the majority of people will pay in this world and economy.

Solved autonomy will be a slow, gradual adder to TSLA and Tesla financials. Sizeable in time yes, but in a LONG time. IMHO.

Remember first that Level 5 is _really_ good.

If people could buy Level 5 as a $15k option take would be huuuuuuge. Not because of offsetting cost by allow it to robotaxi, but because it'd have a enormous impact on commuting and it's a huge enabler for certain groups of people.
It's an ultra-convenient on-demand point-to-point taxi service.

Here's a Linked In article about house prices in the UK for commuting to London, Birmingham and Manchester:

How do commuting times impact house prices and rental costs?

Buying and renting outside of city centres can help homebuyers, landlords and tenants save money. What does rent and house prices look like in commuter towns? The COVID-19 pandemic has changed location preferences for many buyers and tenants.

www.linkedin.com

£449,963 15 minute commute
£512,574 30 minute commute (the sweet spot, presumably because of housing size and green space)
£406,343 45 minute commute.

If you can commute without having to drive, it's much easier to handle a long commute.

Now also consider the benefit for families who move into lower-density areas whose children no longer need a ride.
Or for people who can't drive or who have difficulty driving.

Now, there would likely be some offset to this as affordable L5 should imply excellent cheap L2, but I really think that the impact on housing would be enormous, especially in an era of hybrid and remote working.

Another offset would simply be robotaxi, with people not buying cars. But for the space in which Tesla's currently operating, big uptake.

We don't have L5, therefore it can be hard to understand the impact.

Start by really, really, really considering that with L5 you don't have to drive.
Also really, really, really consider that L5 taxis don't need a driver enabling de facto point-to-point public transportation.

I think more that the key things that investors need to consider is that:
- L4/L5 might not happen
- it might ultimately be very expensive
- particularly that it might not be a monopoly or even oligopoly

While it could be amazing for society, it might never be the license to print money that some assume.

 

[EVCollies]

At some point, Tesla will start Robotaxi 0.1. Driver assisted, limited area and limited users of course.

I would hope this is already happening internally, or am I over optimistic?

 

[Buckminster]

The idea of Tesla licensing their FSD tech has been hypothesized many times here. My question is, is it practical? (I don't have the technical knowledge to have an informed opinion on the subject.) So, I'm asking the bored (SIC) whether it's feasible or not.

Wouldn't Ford, for example, have to use Tesla cameras, HW computer, sensors, brakes, drive train, batteries and steering? Aren't they all connected?

Over the years, I've owned many airplanes. I wanted the latest avionics so I'd do these big upgrades. Never worked out. There was always some problem with one computer talking to another or interference of some kind. I'd suspect that trying to get FSD to work with another manufacturer's product would have the same issues. If anything is different in the system, wouldn't Tesla have to train and certify each unique installation for safety?

 

[Buckminster]

They are all connected but other than the Computer none of them have to be made by Tesla or provided by Tesla or designed by Tesla.

They can use whatever camera supplier they want, brakes, drive train, batteries can all vary.

Just with Tesla, FSD is already working with dozens of different battery pack and dozens of different motor configurations (hundreds of combinations of those). Doesn't really matter to FSD.

 

[Buckminster]

I'll again repeat my FSD/robotaxi thesis.

It doesn't matter if Waymo and Cruise are successful. It doesn't even matter if Waymo and Cruise become profitable.

The only thing that matters is whether or not Tesla's FSD is eventually good enough for robotaxis. If Tesla can achieve autonomy in the next 5 to 7 years then Tesla wins every market they enter. Nobody else can achieve Tesla's low cost. Therefore, Tesla can always undercut the prices of Waymo and Cruise.

 

[Buckminster]

Let's Tesla to the mall:

https://twitter.com/x/status/1696631688502354431

 

[Buckminster]

https://twitter.com/x/status/1704740448626205020

 

[Buckminster]

https://twitter.com/x/status/1716775559378968896

 

[Buckminster]

Extra revenue for owners:

https://twitter.com/x/status/1720918522082078752

 

[Buckminster]

https://twitter.com/x/status/1721035991538880768

 

[Buckminster]

It’s def inference, not training, in the car. Memory and bandwidth not in the same galaxy for training. Dr Know-it-all discusses in this vid.

 

[Buckminster]

Day 12: Twelve Drummers Drumming | Robotaxi at Last​

Part of 12 Days of Christmas - Tesla Edition a series (c) by the Artful Dodger, Dec 2023

Over this Yuletide season, I will post a daily installment focusing on Tesla products, past, present, and future (please note that I will express major themes as short-hand bullet points as yule soon sea. Here's the series so far:

Day 01: A Partridge in a Pear Tree | Roadster Proof of Concept
Day 02: 2 Turtle Doves | S/X Fraternal Twins go Mainstream
Day 03: 3 French Hens | Model 3 Bets the Company
Day 04: 4 Calling Birds | Model Y Built at Four Factories
Day 05: Five Golden Rings | Semi Breaks Physiks
Day 06: Six geese a-laying | Megapack To Excel
Day 07: Seven Swans a-Swimming | Cybertourdeforce
Day 08: Eight Maids A-Milking | Model 2 World Car
Day 09: Nine Ladies Dancing | Dojo as a Service[/HEADING]
Day 10: Ten Lords A-Leaping | Teslabot FTW
Day 11: Eleven Pipers Piping | FSD Or Bust

Intro to Part 12: Twelve Drummers Drumming | Robotaxi at Last​

On July 20, 2016, Elon posted a blog on Tesla's website: Master Plan, Part Deux In it, he introduced the concept that Tesla cars could earn money instead of sitting idle for hours each day:

"In short, Master Plan, Part Deux is:"

• Create stunning solar roofs with seamlessly integrated battery storage
• Expand the electric vehicle product line to address all major segments
• Develop a self-driving capability that is 10X safer than manual via massive fleet learning
• Enable your car to make money for you when you aren't using it

Thus was the genesis of the Robotaxi fleet, a.k.a. "Tesla Network" or TN. Let's explore the rationale, the potential solution, and the resistance to that solution. Lastly, we'll touch on the economics and business implications. Are you ready? Let's GO! (shout out to Warren Redlich)

1. The Problem of Car Ownership: "I like to move it, move it"​

• The most fundamental problem with the way we buy and use cars is that they sit mostly unused for 20+ hrs per day. In fact, they don't just sit idle: they use land for parking, and literally depreciate while they sit waiting
• Cars were build to move. As they get more complex and more expensive, it makes simple economic sense to build fewer cars, but to use each more.
• This is the business model behind rentals and ride-hailing, but they both have one achilles heel: the cost of human labor (at the rental agency, or the driver) is still more expensive in many circumstances than letting a car sit unused. Labor is that valuable
• For example, ride-hailing service Uber often charges $1-$2 per mile, plus surge prices. Meanwhile, the cost of the car (even an ICE-burg) is often only $0.50 or half the cost of labor
• So common folks are stuck in an economic trap: overpay for the convenience of car ownership, or overpay for the convenience of hiring a car

Lesson 1: Transporation should be about convenience; instead its a bother

2. Work, Play or Sleep while you Ride - "Enter the Sandman"​

• The obvious solution is cars that drive themselves (see Day 11 for FSD).
• Without a human driver, and with the lower TCO of EVs, ride-hail could be profitable at half-the cost to customers
• If Tesla solves Autonomy, they will have a huge 1st-mover advantage in a winner-takes-most, $10T market (per ARK Invest 2023 whitepaper)
• I have purposely ordered this business line at the end of Tesla's notable series of accomplishments, simple due to the technical challenges, and the likely resistance from luddites (ie: Dan O'Cloudy, Union-owned Gov't)

Lesson 2: Create a solution where more people are happy and productive, more of the time

3. Fighting the Future - "But my Job (...that I hate)"​

• The problem with the solution is that some people lose their jobs, and some lose income. Whether you drive a Cab, or own a Cab company, there will be less "cabbage" in an Automous future
• Unions especially are afraid of the future, and think little about using mafia-like tactics to forestall the inevitable
• Its very likely that some jurisdictions will be slow to adopt, while some may ban autonomus cars altogether because of these irrational fears of economic loss (BTW, its always a net gain to give people back their time)
• However, more forward-looking jursidictions will become the early-adopters, and their postive experiences and obvious competitive advantages will become undeniable over time (what stays in Vegas?)
• I look for places like Las Vegas and Dubai to be among the first cities to approve Tesla robotaxis
• Ride hailing is just the first application: EV Class 8 transport trucks like Tesla Semi will lower the cost of logistics for most of the goods people buy via lower operating costs and more productive labor (3 trucks, 1 driver platoon)
• This doesn't mean we have less: it means goods cost less, more money left over for other things

Lesson 3: The future is bright, if you would just open your eyes...

Conclusion:​

Both institutional investors like Cathie Wood's ARK Invest and retail investors like Warren Redlich have a long history of predicting Tesla's value based solely on the future value of the Transportation as a Service (TaaS) business opportunity. This is a nascent market which could be worth $10T by 2030, and it could well boost TSLA shares 10x or more from current levels.

Thank you for your attention, Merry Christmas, and Happy Holidays, Everyone!

Cheers, Lodger

*FIN*

 

[Buckminster]

https://twitter.com/x/status/1749752261092466960

 

[Buckminster]

https://twitter.com/x/status/1758039545562779918

 

[Buckminster]

https://twitter.com/x/status/1762311228532068687

 

[MC3OZ]

I have had some thoughts about how to phase in Robotaxis....
The new Gen3 compact can be a Robotaxi or a car sold to a private owner (who may allow it to be part of the Robotaxi fleet).

Assume a factory can make 5,000 Gen3 compacts per week at a 25% margin.
If we allocated 1,000 per week to the Robotaxi fleet and sell 4,000 the effective margin drops to 5% but the gross margin on this part of the business is still positive.
I assume that it takes a compact Robotaxi 1-2 years to work off it cogs, but revenue compounds from day 1.
Starting in an initial city with say 1,000 new Gen3 compacts, Tesla might be able to convince 1,000 private owners to add their car to the fleet and add an additional 100 per week from off lease cars.
Step 1 is to build the fleet in this city which I assume takes 4 weeks for a total fleet size of 4,000-5,000 Robotaxis. if the fleet grows slowly form there private cars and off lease vehicles can keep it topped up.
If the fleet size grows too large for demand, some cars can simply be allocated to the next city.
The process continues for the first year at the end of the year the fleet is around say 60,000 Robotaxis.
For year 2 Tesla tries to scale production up to 10,000 per week allocated 2,000 per week to the Robotaxi fleet, now adding 1 city per fortnight rather than 1 city per month.
What this looks like in terms of year end fleet size/cities is:-

• Year 1 - 60,000 / 12
• Year 2 - 180,000 / 36
• Year 3 - 300,000 / 60

It is a pretty steep launch, but it doesn't break the bank....

After 3 year, tesla possibly continues the ramp process in this country but perhaps adopts a slightly slower ramp in some additional countries.

Gen3 factories needed to be added in sync with the Robotaxi ramp, and eventually rather than 20% of compact production being allocated to Robotaxis, it might be 50%.

The first 3 years and the first 300,000 is the hard part of the ramp, after that time the product and the software should be dialled in...

In the US if regulatory approval is on a state-by-state basis the rollout will also be on that basis. But working Robotaxis on one state definitely helps with approval in other states.

 

[Buckminster]

I completely disagree. On the contrary, Uber’s financials are the best indicator that robotaxis would be wildly profitable. At their current scale they are running approximately at break even profitability. Tesla Network should be able to provide a superior customer experience while also reducing costs by about 2 to 4x relative to the combined costs for Uber and the drivers.

Key numbers from Uber’s 10-K for 2023:
$138B gross booking revenue (not including driver tips)

of which $37B went to Uber​

and $101B to the Drivers (again, not including tips)​

9.4B trips
$1B operating profit
$4.4B spent on sales and marketing (mostly promo discounts and incentives)
$3.2B spent on R&D (almost as much as Tesla’s $4.0B!)

I could not find any disclosure from Uber on their total payload miles, so I can only estimate average trip distance and average booking price per mile. However other sources estimate about $2-3/mile average. I think that’s pretty accurate. Right now in my local market (a suburb of Seattle) the app tells me a typical 7-mile trip is either $3/mile or $4/mile depending on whether I’m willing to wait longer for a car. This is normal for here but the Seattle area is probably pricier than the global average and it’s rush hour right now.

Estimates for net earnings for labor for drivers vary widely but $10/hr is on the low end of the range. This is after subtraction of all the expenses the driver pays for their vehicle. With approximately 40 mph average speed, this is around $0.25/mile for wages. Therefore, driver labor alone is probably at least 10% of the cost structure for Uber’s economic model, and more likely is closer to 20%. If Uber could keep that as revenue then, all else being equal, they’d have earned roughly $15-30B of profit last year. And they are still growing revenue fast.

But it doesn’t stop there. Many Uber drivers are bad at estimating their actual expenses, especially depreciation. Tesla vehicles have some of the lowest total cost of ownership on the market, and that gets better as the annual usage increases. Tesla’s robotaxi vehicle design in particular is designed to deliver the lowest cost of climate-controlled motor vehicle transport ever. The usage pattern will be more efficient than it is for private Tesla owners as well. Robotaxis will be driven gently, and the fleet average range per vehicle will likely be less than it is for current Teslas, because most of the vehicles will stay local and can drive themselves to charging spots during off-peak times. Gentle driving and reduced battery weight means the tires will last much longer. Also, frequent usage throughout the day means the energy cost of battery temperature conditioning and cabin air conditioning is less in proportion to total energy consumption. Additionally, safe driving and drastically lower collision rates should greatly improve insurance costs. The high usage rate also means the capital cost for the vehicle is less, because the interest / opportunity cost on the money sunk into the vehicle is amortized over greater amount of usage per unit of time. Overall, I think Tesla can reduce lifetime average per-mile vehicle expenses by about 50-70% compared to the average vehicle in the Uber fleet.

Tesla obviously is spending R&D vastly more effectively than Uber is. Uber burned 9% of their revenue on R&D with almost nothing to show for it.

Vertical integration would give Tesla further cost efficiency. They can use their own vehicles and their own servicing and charging infrastructure at cost. No profit margin going to someone else and almost zero transactions costs.

Tesla also should have superior economies of scale. With the ability to offer a better experience at lower prices, they can have greater fleet density. This means an higher ratio of payload miles to empty miles, less idle time, and more. Also the service and maintenance will be much more efficient at centralized, specialized facilities than it is for individual Uber drivers who are managing service and maintenance for a single vehicle.

If Tesla Network had provided those same 9.4 billion trips with robotaxis, they’d probably have earned at least $30 billion instead of $1B. 9.4B trips is a tiny portion of overall road transportation demand. This is why, if Tesla becomes the leading autonomous TaaS provider, they could scale to earning hundreds of billions of dollars per year.

That’s the biggest question in my opinion. How many companies figure out autonomy in the first place? Zero? Tesla only? Five?

Even if there are competitors with viable autonomous driving technology, they still would need to compete on cost, service value, and speed of scaling in the early years. Tesla will probably win on every important category:

• Cost to manufacture
• Cost to service
• Vehicle reliability and longevity
• Vehicle manufacturing volume
• Vehicle energy efficiency

Competition is most likely in urban cores. In suburbs, rural areas, and small cities and towns, economies of scale will probably push the market to a winner-take-all dynamic. In any local market, whoever has the biggest network will have the most efficient network, if only because that means less driving time and distance between the endpoint of one payload trip and the starting point of the next. There’s also a winner-take-all effect with the leading TaaS provider being the default option for most people who don’t want to spend time and effort opening multiple apps to shop around. We already see this with Uber having about triple the market share as Lyft.

 

[nativewolf]

I have had some thoughts about how to phase in Robotaxis....
The new Gen3 compact can be a Robotaxi or a car sold to a private owner (who may allow it to be part of the Robotaxi fleet).

Assume a factory can make 5,000 Gen3 compacts per week at a 25% margin.
If we allocated 1,000 per week to the Robotaxi fleet and sell 4,000 the effective margin drops to 5% but the gross margin on this part of the business is still positive.
I assume that it takes a compact Robotaxi 1-2 years to work off it cogs, but revenue compounds from day 1.
Starting in an initial city with say 1,000 new Gen3 compacts, Tesla might be able to convince 1,000 private owners to add their car to the fleet and add an additional 100 per week from off lease cars.
Step 1 is to build the fleet in this city which I assume takes 4 weeks for a total fleet size of 4,000-5,000 Robotaxis. if the fleet grows slowly form there private cars and off lease vehicles can keep it topped up.
If the fleet size grows too large for demand, some cars can simply be allocated to the next city.
The process continues for the first year at the end of the year the fleet is around say 60,000 Robotaxis.
For year 2 Tesla tries to scale production up to 10,000 per week allocated 2,000 per week to the Robotaxi fleet, now adding 1 city per fortnight rather than 1 city per month.
What this looks like in terms of year end fleet size/cities is:-

• Year 1 - 60,000 / 12
• Year 2 - 180,000 / 36
• Year 3 - 300,000 / 60

It is a pretty steep launch, but it doesn't break the bank....

After 3 year, tesla possibly continues the ramp process in this country but perhaps adopts a slightly slower ramp in some additional countries.

Gen3 factories needed to be added in sync with the Robotaxi ramp, and eventually rather than 20% of compact production being allocated to Robotaxis, it might be 50%.

The first 3 years and the first 300,000 is the hard part of the ramp, after that time the product and the software should be dialled in...

In the US if regulatory approval is on a state-by-state basis the rollout will also be on that basis. But working Robotaxis on one state definitely helps with approval in other states.

So far the state in the USA that really matters is California. It doesn't give a flip what other states are doing, most states have no regs (but they will come). Tesla has done nothing with the CA DMV so expect a long delay there. This would impact 2 of the most important global profit markets for TAAS, LA and San Fran metro region. Just eyes wide open.