A privately owned car is useless battery capacity that is sitting there not driving and fully charged for most of the time while you are sleeping, working, or whatever (i.e. not driving). An autonomous car owned by a company that sells rides is very different. It will be ferrying around people whenever it can. You need a lot less of those to satisfy demand.
Tesla will be selling lots of vehicles and batteries before we get there of course. But it would be a mistake to think of the market in terms of how things currently work. Which is the mistake this article makes.
That’s not necessarily true in the future. One idea I’ve heard is to use charging cars for energy grid balancing.
If Tesla said that it would cost $5T instead of $10T to electrify the world because of FSD the world would laugh and dismiss the argument wholesale. 42M is the only defensible number they can use in their public calculations.
That part of the document is calculating the global annual battery demand once we're fully electrified. The chart says that 42M compact cars are sold annually by all vendors across all countries, and that if electrified, those 42M compact cars will have an average battery pack size of 53kWh, leading to an annual demand for 36TWh of batteries for that segment.
Telsa obviously doesn't believe they're going to get 100% of the market, they don't believe they're going to sell 42M compact cars per year. The 53kWh figure is no more meaningful.
Heck, that TBD in the table doesn't even promise that Tesla will even enter the segment, internet commentary to the contrary.