- a guaranteed purchase of 10k cars will get a significant discount over retail price
- not paying dealer overhead
- buying gas (or electricity) in bulk will be cheaper than retail price
- in-house repairs and maintenance will be cheaper
- a lot of people buy on credit, which is expensive. A fleet operator will be able to get cheaper funding (e.g. free by doing an IPO) and eventually fund purchases fully from profit
- an average customer is influenced by marketing and therefore over-pays for non-essential items. A fleet buyer will be ruthless about maintaining best-bang-for-the-buck ratio.
In the longer term fleet operator will transition fully to electric, build its own electricity generation capacity (especially in sunny states like texas or ca, where solar is really cheap), use data from past repairs to drive up reliability and durability of the cars etc.
There's plenty of other things to break of course, but when the business model includes selling mobility the incentives are aligned toward durability and reliability (versus selling cars, which incentivizes planned obsolescence). It's worth putting in that extra 5% to make the car last twice as long.
That means driving as much as a NYC taxi (70K mi/year) instead of as much as a normal car owner.
The per-mile cost also includes the capital outlay for the car itself. In other words, it might not make sense for you to pay $40,000 for a new electric car, but it could make sense for three people to split that, and drive each persons per-mile cost down.
All of this is in reference to how fast the switch over could happen, and is why it might be much faster than any of us anticipate.
The obvious answer - dividing the cost of the car over more miles - doesn't seem right to me. Driving more miles doesn't generally increase the useful mileage lifespan of cars, it just wears them out more quickly.