I have looked extensively at traffic patterns. Extensively, it fascinated me. The people driving to and from metro stations, the people dropping kids off at schools, the landscapers, etc. I actually spent days looking and then decided it didn't matter. Most people that drove to work, park, then they don't move the cars. Most commuters, most days. That's well documented. So, the people on the road in the middle of the day are usually not commuters. Some are going to meetings but they are people doing something (fedex, delivery XYZ, service XYZ, kids, tourists, etc). Again it does not matter. Take Seattle again. 350k workers downtown. 25% commute in single use cars trips. 80k or so cars moving in and then leaving once a day. I use Seattle because my own metro region is so F'd it makes the numbers look worse for RT.
The RT picks someone up in suburbs and takes them downtown. Between 8- 9am the RT, 80K of them, are not parked but sitting in the roads causing even worse congestion. Where do they go? There are not 80k riders needing to go somewhere in downtown Seattle at 9am. Uber handles that 9am traffic in Seattle with a couple of thousand cars. That is the market downtown for people that didn't want to go around in a rental or in their own car. So you have a gap of 78k or so cars. That's a lot of traffic. A lot. Where is the business? It isn't there at 8-9 am. Cars are moving but it isn't 78k worth of cars moving. They are also not commuting regularly. People are working from home and a certain amount of that never comes back. That might help the RT use case I am just not sure. It means you have to have a huge amount of vehicles stashed to handle surges on Tues-Thursday.
If the commuters cars are the RT fleet then there are 80k chasing 2k worth of work. Does everybodys car get to take 2 fares/week? I don't see how that is even worth signing up in the system (the intercity midday fares in Uber are very low, 2-4 miles 10-15 mins- this is all about parking and convenience). An uber driver may get 15 such rides between rush hours. The pofitiable rides for Uber are bar hopping rides starting at 5pm and ending at 11pm but it's really not a big number. From 8pm to 6am there isn't much traffic except for this work and occasional airport work.
The greatest societal good I see from working RT is that it brings new societal interactions to elderly, handicapped, and substance abusers. Can they use an app? I guess we can make sure they are accessible. will the rt be handicapped accessible? Maybe Uber drivers will still be needed to move wheelchairs or ? Anyhow, lots of good but I am not sure how much profit.
Do your own analysis. Figure out how to move the commuters and then how to move the other people. Carve out all the people with work related vehicles. Go stand at a street corner and just count each car at a corner. Easy RT replacement or not.
Then here is another issue. We need to park and charge a minimum of 75k cars.
Will Tesla invest in 75k cars to handle Seattle transport market? If they had invested in 2018 they'd have needed 100k cars but people are teleworking so the number decreased across the USA. What would they have done with the extra 25k cars?
I think Waymo is on the right track. Gut Uber and grab all the dense Urban traffic and as demand slowly increases take the profits and expand. If Waymo is first you'll have 50k vehicles making profit. Waymo could be serving the 5 big CA cities by end of 2025 and that would be 5/20 in the USA. Obviously they add Austin and Phoenix and then some others. We'll see. They are being very methodical.
Google can better leverage the ad dollar spend associated with a captive ride, they'll know if you buy pizzas or pho on the way home and can hit you with ads or just to extend trip (OR NOT). For this reason I've always thought Google was in a good position. Also that Ubers data has some value.
Why focus only on commuters? That's a major part of demand, but far from the only portion that matters. The transportation demand curves clearly show that. A work-related vehicle is not exactly a robotaxi but if it's getting high usage for a commercial or government application, then it's similar economically. Also not everyone is on the same shift time, especially in a market like Seattle. Rush hour is not just 8-9AM and 5-6PM. It's more like 6-10AM and 3-7PM, and the evening rush hour largely goes in all directions, not just away from downtown. Then there's lunchtime trips/deliveries, midday errands, picking up sick kids from daycare/school, stay at home parents, second-shift workers, medical appointments, and more. Also the robotaxi is most competitive not against private cars for suburbanites, but rather against trains and busses for city dwellers. Robovans would be a more efficient and effective version of traditional mass transit solutions. At perhaps 8-16 passengers worth of capacity, they can be big enough to get decent economies of scale, but small enough to offer reasonably fast service without too many stops. Busses don't have as good of a balance because the overhead costs of driver wages and engine maintenance favor a larger vehicle size.
Since the average Seattle commute time is about 30 min, that means taxis could squeeze in around 3 or 4 rush-hour commutes each morning and each evening, and that’s conservatively assuming most of those trips require a hour-long round trip to allow the robotaxi to return to the suburbs to pick up the next rider. In reality, a lot of people, especially in the urban core, are commuting not radially in and out of the downtown center but rather from one part of town to another.
Parking and charging is easily solvable. It just takes some infrastructure investment. Tesla could repurpose vacant lots of defunct suburban shopping mall lots, for example.
You said, "That is the market downtown for people that didn't want to go around in a rental or in their own car." Too much of this is static analysis. Robotaxi is not just an Uber or taxi without a driver. That's effectively all Waymo is accomplishing, but they're not capitalizing on the whole opportunity because their current technological capability does not allow for that and because they don't design and manufacture their own vehicles.
1) "The market" is a combination of supply and demand curves. Robotaxis represent an increase in the supply of transportation, meaning that high quantities can be delivered at lower prices. The new market equilibrium would have higher quantity demanded. Ubers at peak hours often cost $3/mile or more. As I write this, it's 7AM here, before the peak of morning rush hour, on a Monday. Right now the price of an Uber ride 17 miles from here to downtown is $48. After tip, that comes out to $3/mile. I wouldn't even consider it at that price. But for $10 I might.
2) As many people have already expressed on this forum, not having a driver is more than just a cost savings. It's also higher value. People want the privacy and lack of awkward silence or small talk expectations. People want to be able to listen to whatever music or entertainment they want
3) Dedicated robotaxis can exploit the economics of high efficiency usage to offer a more luxurious and premium experience. They will have the budget to do so because of the high utilization rate and long vehicle lifetime compared with the average ICE/hybrid Uber. I
posted about this a couple weeks ago. There could be a big screen for entertainment, premium audio system, nicer seats, and so on. It can be more like a miniature living room or like a first-class cabin on a long-haul airplane.
4) Big wildcard: Boring co tunnels. If this works out, robotaxis will offer significantly higher speed than any Uber or taxi ever could.