The math works out nicely if you look at it this way, but the problem with uber is that it is people intensive. Their core asset are their drivers and not their technology. To maintain that driver supply they are paying out their drivers quite handsomely. Just to add I am currently in the largest south east asian market and there are 3 major competitors here fighting for the 250M+ population.
If you cut out those subsidies, you cut out the advertising, those numbers would go into a downward spiral. Drivers would go to the competition who keep subsidizing, there would be less cars as a result, less cars means a less useful service and users would eventually leave, which would mean THE END.
I personally think the core to this issue with Uber and all other ride hailing businesses is that the behavior they are promoting is far from what people deem as normal. A few years back , letting some random person into your car and taking them somewhere was the beginnings of many horror movies either for the driver who would be the victim or visa-versa (it would just depend on the title of the horror movie). In this day and age i still think its not normal to let a random person in your car, but with enough money to do it people will allow for this behavior. I mean if you were not getting paid enough would you let someone into your car randomly and take them where they wanted to go.
With this in mind uber must shell out the cash to allow drivers to want to loan out their cars for others to get in and also give up their time driving around waiting for a passenger to hail them and since they are not the only service like this in each country the cost gets exponentially hire when they compete for price hence why the crazy fund raising required by all businesses of this type in each specific region.
Very different to online marketplaces where the assets are the merchants within the marketplace who sell. In this scenario merchants sell, its what they have always done and is normal behaviour to them so having a channel to sell their goods to a wider audience is always going to be a better option and merchants hence dont need $$$ to keep them within the market place. Once you get them on board you dont have to deal with them potentially leaving, they can sell on your platform or any other platform it just comes down to driving traffic.
Even though with ride hailing you can be part of many different ride hailing services at once, if you are driving for one you are not driving for the other at that moment so their is solid competition for drivers time there and drivers will go with who pays the most which then comes in the for of subsidies. You let those subsidies go you loose your drivers, you loose your users and you loose the market share, then you loose.
Assets of ride hailing services are humans, humans change and get restless and always looking for something better, maintaining that asset happy without lots of cash (specially when there is other competitors doing the exact same thing) with more or less the same technology is going to require lots of cash.
Im curious to see where all this will end, for now im enjoying my crazily subsidized rides that has helped me save so much money in transportation, so im not complaining , but where this business will go eventually and where the regional players that compete with uber on a regional level are all going to end up is going to be dam interesting to see ... im excited :)