Robotaxi : The business of competing with human drivers

[EVNow]

I've seen various business cases for Robotaxis - but the main competition are human drivers. So, how much can RT make when competing with human drivers ?

Let us say the minimum wage is $15 an hour. A lot of Uber/Lyft drivers actually make something like or even less that once all costs are accounted for.

All other costs of operating a RT is no different than operating a taxi service. Utilization of a taxi can be as high as RT when using drivers in shifts. Both RT & Taxis need to be cleaned, serviced, charged etc. The only difference is who is driving - a human driver that needs to be paid $15 an hour or the software. For the time being we can assume a taxi and an RT costs the same in terms of hardware.

Let us say RT can be used for 20 hours a day - leaving 4 hours for maintenance / charging. So, per day we get

20x15= $300 a day or $9,000 a month or about $100,000 a year.

So, an RT can replace human drivers who need to be paid $100k per year (min wage). That is really the "value" of RT.

This is the reason, in other countries, RT may be of different value. In India it might be just $10k per year (so, clearly not worth it) or some $20k in China - so is it really worth it ? At $100k a year on investment of say $50k for the RT, clearly RT is a sound business.

Ofcourse the question of demand, utilization etc will dictate the actual revenue - just as it would for a Uber driver or taxi operator.

"Tesla Model 3, prodotto il primo esemplare. Ecco com'è la berlina compatta di Musk" by automobileitalia is licensed under CC BY 2.0.
Admin note: Image added for Blog Feed thumbnail

 

[spacecoin]

Yeah it's not a slam dunk by any means, but from what I've read 90% or the revenue is in large metros, so clearly geofencing isn't a problem. It also goes hand in hand with the fact that depots and maintenance crews that are fairly fixed geo-wise.

I like the reasoning around worker costs in different parts of the world.

 

[RTPEV]

Ofcourse the question of demand, utilization etc will dictate the actual revenue - just as it would for a Uber driver or taxi operator.

I really like the question that you've asked here.

My stumbling block with respect to robotaxis, however, is not on the demand or utilization side, but the supply side, and I think the comparison to human drivers breaks down when we consider that.

It seems the business cases that people envision for robotaxis are in fact based on comparing to existing rideshare/taxi revenue, which as you correctly point out, assume having to pay a human driver. Take the human out of the equation and voila, you have a strong business case. Or do you?

What I think would (will?) happen is that the market would then be flooded with robotaxis, dramatically increasing the supply, which in turn will cause the revenue per ride mile to fall. Of course an equilibrium will be found, because at a certain price point it doesn't make any sense to invest in a robotaxi capable vehicle, but for sure the cost per ride mile will wind up being at a point I think will be quite a bit less than today's human driver ride mile.

I'm not sure those investing in robotaxi technology, or are betting on the future value of robotaxis are taking that into account. Either that, or I'm just wrong.

The other option is that people really do give up their cars so that demand goes up correspondingly with the increase in supply. I suppose this could happen, but outside of dense urban environments, I just can't see that happening. There are too many use cases for personal ownership of cars that we've become accustomed to (at least in suburban and rural America) that I don't see what being the case any time soon.

 

[EVNow]

What I think would (will?) happen is that the market would then be flooded with robotaxis, dramatically increasing the supply, which in turn will cause the revenue per ride mile to fall.

That would happen only if Robotaxis get commoditized. That is unlikely - in the short term. Esp. considering regulations. I expect few RT players who basically compete with human drivers (atleast in the first few years). That is why we need to compare to human drivers.

I'm not sure those investing in robotaxi technology, or are betting on the future value of robotaxis are taking that into account. Either that, or I'm just wrong.

They are expecting to be the first to expand and take a big market share - and continue to operate generating healthy margins, even if reduced after stabilization.

 

[Max Spaghetti]

The first mover advantage will likely only really apply to the first company able to scale rapidly. Uber did this, spending way more than it was earning in the early years - presumably trying to get in everywhere around the world before anyone else. But now that Uber has competition, the supply/demand equation applies, with a lower pricing limit based on how little drivers are prepared to drive for (and Uber's cut).

RT's will remove the driver pay from the equation, but someone will still have to be paid to clean and charge the vehicles (unless they can work out self-charging, but cleaning will still be necessary). Currently, drivers clean and charge/fuel their own vehicles, so you are not paying extra for that. But for an RT, that is an extra cost, so you are not removing the 20 hours/day of wages fully. Also, human nature being what it is, I wonder how much respect people will have towards a robo-taxi rather than a car that is owned by the driver sitting in there with you (by which I mean more likely to leave the car a mess or even wilfully damage it). Cleaning may be significant. I wonder how much trouble Waymo has with that?

Also, the 20 hours/day of wages is not really applicable. Most taxi/rideshare drivers are paid by the trip, not the hour. During off-peak periods, they may be sitting around a lot unpaid. They are also not paid while travelling to a job empty.

But once the RT supply/demand equation takes place with a lot of competition, the price of a ride to the consumer will likely end up being the cost of the service (charging, depreciation, maintenance, cleaning) plus whatever small margin the cheapest competitor is prepared to endure. No-one will be raking in $100,000 per year per vehicle, because someone else will rapidly undercut them. So RT's will likely end up being cheaper than taxis/rideshare, which may rapidly kill off that industry to all but those who might need or want human drivers (e.g. elderly, disabled, robo-phobes, etc).

Another interesting question is which business model will emerge?

1. The Waymo model, where they design, build and own the vehicles (or modify off the shelf vehicles); design and own their own software; design and own their own app; employ staff to do the cleaning/charging, and keep all the profit? or

2. The Uber model, where they design and own an app, and just sign up individual owners of RT's who may have purchased the vehicle from any number of manufacturers - the only proviso would need to be that the RT would need to communicate with the Uber server - so a version of the app onboard somehow (to replace the phone that the Uber driver carries). When the RT needs cleaning or charging, it will either autonomously return to the owner, or to a designated service provider of charging/cleaning, which the owner pays for.

Some have hinted that Tesla wants to own its own fleet of RT's (the Waymo model). But I think a lot of Tesla owners would like their personal cars to be capable of L4 driving, so that they can be chauffeured around in their own vehicle, and potentially (or not) send that vehicle out to earn more money when they don't need it (the Uber model).

 

[primedive]

As an N=1 viewpoint, I personally would pay more for Robotaxi than for Uber, even if due to economics the pricing is likely to be less. Some reasons why:

* Standardization of quality and maintenance. No more getting picked up by beater Toyotas or Hondas.
* No bull*sugar* about the driver rating you poorly because you have children, suitcases, he's in a bad mood and so on.
* No need to tip. Uber is at minimum 20% more expensive due to this tipping bull*sugar*. Tipping in general outside of narrowly defined niches is a cancer on society IMO, but having it used as a subsidy for their business is pernicious in particular.
* No dangerous driving, such as when one of my drivers took an ULT with an 18-wheeler literally less than 1 second away from smashing into our side.
* Ability to converse loudly, choose music, choose climate, and so on.

There are so many advantages with the robotaxi model that you just don't get with your regular human driver.

 

[LostVector]

I've seen various business cases for Robotaxis - but the main competition are human drivers. So, how much can RT make when competing with human drivers ?

Let us say the minimum wage is $15 an hour. A lot of Uber/Lyft drivers actually make something like or even less that once all costs are accounted for.

All other costs of operating a RT is no different than operating a taxi service. Utilization of a taxi can be as high as RT when using drivers in shifts. Both RT & Taxis need to be cleaned, serviced, charged etc. The only difference is who is driving - a human driver that needs to be paid $15 an hour or the software. For the time being we can assume a taxi and an RT costs the same in terms of hardware.

Let us say RT can be used for 20 hours a day - leaving 4 hours for maintenance / charging. So, per day we get

20x15= $300 a day or $9,000 a month or about $100,000 a year.

So, an RT can replace human drivers who need to be paid $100k per year (min wage). That is really the "value" of RT.

This is the reason, in other countries, RT may be of different value. In India it might be just $10k per year (so, clearly not worth it) or some $20k in China - so is it really worth it ? At $100k a year on investment of say $50k for the RT, clearly RT is a sound business.

Ofcourse the question of demand, utilization etc will dictate the actual revenue - just as it would for a Uber driver or taxi operator.

View attachment 1047593
"Tesla Model 3, prodotto il primo esemplare. Ecco com'è la berlina compatta di Musk" by automobileitalia is licensed under CC BY 2.0.
Admin note: Image added for Blog Feed thumbnail

The X factor in all of this is how defensible / non-reproducible a robotaxi is. If self driving ends up being easy to replicate or copy due to general advances in AI, no robotaxi business will sustain unusually high margins.

 

[electrons4u]

... But I think a lot of Tesla owners would like their personal cars to be capable of L4 driving, so that they can be chauffeured around in their own vehicle, and potentially (or not) send that vehicle out to earn more money when they don't need it (the Uber model).

Tesla owners might not want to send out their personal car to earn money because what happens when it picks up some drunk people at a bar (who *do* need a ride) and one of them pukes all over the seat(s)? Who's going to clean it? You might be able to tell the car to not do bar pick-ups. Or maybe people buy an old used Tesla to serve as the barf bus.

 

[KerrySkates]

Some have hinted that Tesla wants to own its own fleet of RT's (the Waymo model). But I think a lot of Tesla owners would like their personal cars to be capable of L4 driving, so that they can be chauffeured around in their own vehicle, and potentially (or not) send that vehicle out to earn more money when they don't need it (the Uber model).

Musk said on the most recent earnings call that Tesla would have their own fleet and that owners could choose to supplement that fleet with their personal vehicles.

 

[spacecoin]

Musk said on the most recent earnings call that Tesla would have their own fleet and that owners could choose to supplement that fleet with their personal vehicles.

I doubt that's going to happen. It's basically reverse car-rental. If the car isn't returned to you in the same condition, who deals with that? Damaged exterior/interior? How do you prove it? I wouldn't rent out my car if there are no guarantees about that. In fact I probably wouldn't rent out it regardless.

On another note, Baidu are launching their sub-$30k robotaxi in Wuhan with all required sensors (camera, radar, lidar) in the summer.

 

[KerrySkates]

I doubt that's going to happen. It's basically reverse car-rental. If the car isn't returned to you in the same condition, who deals with that? Damaged exterior/interior? How do you prove it? I wouldn't rent out my car if there are no guarantees about that. In fact I probably wouldn't rent out it regardless.

On another note, Baidu are launching their sub-$30k robotaxi in Wuhan with all required sensors (camera, radar, lidar) in the summer.

My guess is that individual owners will need to pay for some type of taxi-like insurance on their vehicles to cover these costs. Ideally the cameras around the vehicle will have continuous capture so they don’t miss anything that is in their view. Of course, violators could just cover the cameras so it’s not going to catch all violators.

Tesla and Individuals will have the same risk. I doubt many owners will find it worth the time, effort and cost and instead this will be people who build their own mini-fleets, like a taxi business.

It would take some time for the industry to figure out the details. Early adopters will have to find out through trial and error. High risk - but is there also a high reward to jump on board early? Idk.

I don’t think any of this is happening any time in the near future anyway. FSD isn’t anywhere near ready. Robotaxi/Cybercab is a pipe dream for now.

 

[BigNick]

My guess is that individual owners will need to pay for some type of taxi-like insurance on their vehicles to cover these costs.

There goes your profit.

Taxicabs usually carry only liability insurance because comp/collision is WAY too expensive. Most cabs are ten-year-old plus throwaway cars.

Uber/Lyft/etc drivers don't usually carry proper insurance for taxicab duty despite their effective usage being the same. Otherwise they wouldn't be able to afford the comp/collision coverage required by their financing.

 

[ChooseFreedom]

I really like the question that you've asked here.

My stumbling block with respect to robotaxis, however, is not on the demand or utilization side, but the supply side, and I think the comparison to human drivers breaks down when we consider that.

It seems the business cases that people envision for robotaxis are in fact based on comparing to existing rideshare/taxi revenue, which as you correctly point out, assume having to pay a human driver. Take the human out of the equation and voila, you have a strong business case. Or do you?

What I think would (will?) happen is that the market would then be flooded with robotaxis, dramatically increasing the supply, which in turn will cause the revenue per ride mile to fall. Of course an equilibrium will be found, because at a certain price point it doesn't make any sense to invest in a robotaxi capable vehicle, but for sure the cost per ride mile will wind up being at a point I think will be quite a bit less than today's human driver ride mile.

I'm not sure those investing in robotaxi technology, or are betting on the future value of robotaxis are taking that into account. Either that, or I'm just wrong.

The other option is that people really do give up their cars so that demand goes up correspondingly with the increase in supply. I suppose this could happen, but outside of dense urban environments, I just can't see that happening. There are too many use cases for personal ownership of cars that we've become accustomed to (at least in suburban and rural America) that I don't see what being the case any time soon.

Correct. Plenty of variables that are in play here that are always in flux:

RT quantities in an area, compared to taxis.
Supply/Demand for rides
Pricing

And if left alone, the free market will find a balance.
But Govs can never leave anything alone, especially if they see a way to get their cut.


Another interesting dynamic I'm curious to see play out will be the reactions of the cab companies, and even more so, the drivers unions.
What happens when/if a cab company wants to add RT's to their fleet?
What legal action will a union take against those cab companies, or even a new startup RT-only company?
Does NYC have a cab driver's union? If so, they probably have quite a bit of political influence in city gov policies. They could use that to try and get RT's banned, or put severely limiting parameters on them.
Didn't they already try that with Lyft/Uber? But because of the human impact, this is vastly different.
Lyft/Uber, for the majority, affect humans that just want to drive part time typically, and didn't want to work for a cab company, so the social sympathies were with the David drivers vs the Goliath cab companies.
RT's flip that sentiment. It then becomes cab drivers vs robots.
Such an interesting social experiment!

 

[kabin]

On another note, Baidu are launching their sub-$30k robotaxi in Wuhan with all required sensors (camera, radar, lidar) in the summer.

Speaking of which. There was recent talk of a Tesla and Baidu working together on RTs. Probably a low production high cost version, but I was surprised to see Baidu already has a working RT. CNBC's Eunice Yoon took one for a ride.

https://twitter.com/x/status/1791109157631791582

 

[kabin]

Uber/Lyft/etc drivers don't usually carry proper insurance for taxicab duty despite their effective usage being the same. Otherwise they wouldn't be able to afford the comp/collision coverage required by their financing.

I would hope Uber and Lyft would closely monitor driver insurance policies before allowing them to operate.

 

[enemji]

I've seen various business cases for Robotaxis - but the main competition are human drivers. So, how much can RT make when competing with human drivers ?

Let us say the minimum wage is $15 an hour. A lot of Uber/Lyft drivers actually make something like or even less that once all costs are accounted for.

All other costs of operating a RT is no different than operating a taxi service. Utilization of a taxi can be as high as RT when using drivers in shifts. Both RT & Taxis need to be cleaned, serviced, charged etc. The only difference is who is driving - a human driver that needs to be paid $15 an hour or the software. For the time being we can assume a taxi and an RT costs the same in terms of hardware.

Let us say RT can be used for 20 hours a day - leaving 4 hours for maintenance / charging. So, per day we get

20x15= $300 a day or $9,000 a month or about $100,000 a year.

So, an RT can replace human drivers who need to be paid $100k per year (min wage). That is really the "value" of RT.

This is the reason, in other countries, RT may be of different value. In India it might be just $10k per year (so, clearly not worth it) or some $20k in China - so is it really worth it ? At $100k a year on investment of say $50k for the RT, clearly RT is a sound business.

Ofcourse the question of demand, utilization etc will dictate the actual revenue - just as it would for a Uber driver or taxi operator.

View attachment 1047593
"Tesla Model 3, prodotto il primo esemplare. Ecco com'è la berlina compatta di Musk" by automobileitalia is licensed under CC BY 2.0.
Admin note: Image added for Blog Feed thumbnail

Considering all things equal, the difference is between the cost of having a human driver vs a team of researchers, scientists, engineers who develop and maintain the AV software.

 

[spacecoin]

Speaking of which. There was recent talk of a Tesla and Baidu working together on RTs. Probably a low production high cost version, but I was surprised to see Baidu already has a working RT. CNBC's Eunice Yoon took one for a ride.

There is zero chance of that happening. Western "media" got that wrong. Baidu said "Baidu will seek cooperation opportunity with Tesla on Robotaxi based on Tesla’s specific need and timetable."

It's a misconception that they would use Tesla tech, it's the other way around if anything.

They also recently said that they believe they are at least 3-5 years ahead of anything Tesla has.

https://twitter.com/x/status/1790954051280220367

 

[EVNow]

RT's will remove the driver pay from the equation, but someone will still have to be paid to clean and charge the vehicles (unless they can work out self-charging, but cleaning will still be necessary). Currently, drivers clean and charge/fuel their own vehicles, so you are not paying extra for that. But for an RT, that is an extra cost, so you are not removing the 20 hours/day of wages fully. Also, human nature being what it is, I wonder how much respect people will have towards a robo-taxi rather than a car that is owned by the driver sitting in there with you (by which I mean more likely to leave the car a mess or even wilfully damage it). Cleaning may be significant. I wonder how much trouble Waymo has with that?

Also, the 20 hours/day of wages is not really applicable. Most taxi/rideshare drivers are paid by the trip, not the hour. During off-peak periods, they may be sitting around a lot unpaid. They are also not paid while travelling to a job empty.

Basically - if you operate a taxi company efficiently, you would have drivers on shifts and separate cleaning staff. So, it finally comes down to driver pay.

In current taxi business, generally the drivers are leasing the taxi and paying some fee (or a % of revenue). They don't really get paid by the hour or miles or for the # of hours worked. But, for taxi driving to be a viable job, it needs to pay about minimum wage, atleast. Otherwise the driver will just flip burgers for min wage. According to some websites, drivers get about $20 an hour.

So - in other words - what you are doing in RT is replacing the labor of a human, so it is worth atleast minimum wage.

But once the RT supply/demand equation takes place with a lot of competition, the price of a ride to the consumer will likely end up being the cost of the service (charging, depreciation, maintenance, cleaning) plus whatever small margin the cheapest competitor is prepared to endure. No-one will be raking in $100,000 per year per vehicle, because someone else will rapidly undercut them.

That is eco 101.

As I said, this is assuming RT is commoditized. But that will take a LOOOOONG time. Even then, there are network effects, regulations that will limit the numbers and companies will continue to get huge margins like Apple does today. Just a question of how you differentiate yourself from the commodity. Anyway, that is all far fetched at this point. Initially there will only be just a couple or handful of operators in each market, competing with human drivers.

 

[EVNow]

Considering all things equal, the difference is between the cost of having a human driver vs a team of researchers, scientists, engineers who develop and maintain the AV software.

The basic idea of software is - the marginal cost is almost zero. So, you pay the engineers the same amount whether you have 100 RTs or 100,000. So, a very large fixed cost, but a very low marginal cost.

 

[EVNow]

Lyft/Uber, for the majority, affect humans that just want to drive part time typically, and didn't want to work for a cab company, so the social sympathies were with the David drivers vs the Goliath cab companies.
RT's flip that sentiment. It then becomes cab drivers vs robots.
Such an interesting social experiment!

Yes - Uber/Lyft vs Taxis were basically one set of drivers vs the other set. In general people want more choice (and Uber/Lyft is so much easier to hail) - so will always vote for Uber/Lyft.

But when it comes to human vs robot drivers, it will take time to replace humans. But will eventually happen - it has happened in all other occupations where it was possible.