… that deliberately takes routes which maximise your exposure to its owners advertising, services and products.

... unless they're interested in taking in a loss for each vehicle or only selling to the richest people in the world.

The business model for the self-driving car is a computer-scheduled, dynamically-provisioned taxi service cheap enough that nobody needs to own a car in the city. Because self-driving cars don't need to park (they can just move on to the next assignment or deprovision), if they catch on then in-city parking decks can be redeveloped into business space or park space.

You forget transportation of goods. Pretty much all over-land cargo in the U.S. and most of that elsewhere (less so, because trains) could be transported by self-driving cars/drones. Not sure if that is larger or smaller than transportation of people at this point, but it's still an astronomically huge market.

US is actually pretty big in freight trains.

Mostly because, freight and passenger trains interfere badly. No passengers = easier freight. A big wave of deregulation of the market a few decades ago helped, too. The Economist ran a special about the US market a few years ago.

Transport is something like 30% of the world economy isn't it?