The comparison was with today's costs for Uber and if you think a bit beyond the obvious, the leading CaaS providers can produce their own fuel. Industrial consumers already pay about half as much for electricity as the public but costs can be optimized further.
The CaaS provider doesn't need a 3rd party insurer. Tesla is actually already moving towards a single price for the car , maintenance and insurance - with or without 3rd parties involved for now.
Something i did not not mention is that the CaaS provider will design their own vehicles, aimed at convenience and optimized for operating costs - there is a lot that can be done there.
I would put a question mark on "electric powertrain" because the manually driven Uber car will be electric too.
Also, for the record – the insurance premium will be lower for the robot taxi, since it has fewer passengers onboard to get hurt, for example, byt maybe that's not a big difference since the manually driven Uber will have crasch avoidance too.
For a car as a service the vast majority of the costs are operating costs. When a vehicle travels up to 500k km in a single year, the hardware costs fade into irrelevance. Cost reduction drivers - electric powertrain (higher efficiency and much cheaper fuel), no human driver, much higher utilization than current services, scale- scale is important , future providers will be much much bigger than Uber can ever be. Or think about it this way, 4 billion people paying on average 125$ per month or 1500$ per year (half of them would only spend 50$per month, that's why 4B customers and ASP so low) for a TAM of 6 trillion a year. That's where the world needs to be by 2030. The 2-3 leading global providers, each with 20-30% share, will be kings...
Recurring cost of the use of the robotic vehicle is likely to be lower; however, the initial cost of the vehicle, plus all of the sensors, not to mention recovery of the software development and all of the testing that the comments (so far) are waiting for, is likely to drive up the final price tag. I'm not sure that the cost will be less than Uber.
(For the record, I am not an Uber driver or user.)
Why would you choose a human driven taxi over a robotic one?
Excellent points. Not to mention lower chance of Rape/Assult...
Unfortunately, it's likely to be a while before I see (or trust) autonomous cars around here. I notice that they seem to always run their tests in places with nice weather. I'm not going to trust them here until I see sucessful tests in places with snow, ice and hills.
"Why would anyone choose a robotic taxi over a human-driven car?"
It's amazing how many analysts not only lack vision but are not even mildly informed.
It's not just cheaper than Uber,it's cheaper than the gas your own car uses (in areas like Europe where gas is not cheap) and by 2025 is cheaper than a bus. It's safer too and at maturity it provides a lot more convenience than any current vehicle as service providers can go wild with their offering - something nobody can do at reasonable costs today. Cheaper is not the only option, more for the same $ is also an option. Folks that drive a Corolla today can save a few thousands $ per year or they can spend the same and ride in much much nicer vehicles.
As for how folks react to self-driving technology, he should go on Youtube and watch Tesla videos for a few weeks. People adapt very quickly to convenience and by the time such services arrive, many will have some experience with autonomous features.
I think the question the analyst asked in the article is backwards. Why would you choose a human driven taxi over a robotic one? You don't tip a robot, so even if the base rates were the same the robot will be cheaper. You don't have to worry about a robot driver getting distracted by a text message or a fight with his wife, or smells bad because he chain smokes between rides or has nasty BO.
I wouldn't choose a robot driven taxi until the whole 'test' phase is over, and you can buy autonomous cars from most automakers. They'll have to bugs worked out by then, and it should be a lot safer than human drivers since they can pay attention 100% of the time, "see" in the dark/fog/low sunlight, will have awareness of the traffic for miles in any direction so they can route around congestion or accidents.
So while I'm not going to be a guinea pig for Lyft when they have a few hundred pilot cars out debugging their system, once autonomous cars appear for real pretty much every taxi/Uber/Lyft driver loses their job overnight.