In this discussion I have not seen any consideration of the avoided cost of advertising. Tesla, unless I'm mistaken, spends zero on advertising, whereas other manufacturers spend on the order of $500 per car sold.
$500 can buy you a whole lot of electricity. At reasonable rates that's about 5 MWh, enough to power a Model S for approximately 25,000 km (15,000 miles).
But it's more reasonable to compare the cost of advertising with the cost of building the Supercharger network. Right now there are 500 Supercharger stations. It's pretty tough to come up with an average installation cost, since some are 2-stall units and some are 10-stall, but I think a fair estimate is about $200k per. So Tesla has so far spent about $100 million to build the Supercharger network. They have also sold roughly 100,000 cars. So that's about $1000 per car. But this rate of Supercharger installation will, I think, not keep up with sales in the long run, as most Supercharger sites have very low usage currently. They have had to spend a lot to provide geographic coverage.
So, to within some reasonably large margin for error, over the long term Tesla can provide the Supercharger network essentially at ZERO COST relative to what other automakers spend on advertising.
The question then becomes "what about the cost of energy?"
If you take the $2000 charged for an S60 to be Supercharger-enabled as the price charged for all Model Ss built, then you're talking about a break-even point of about 100,000 km (62,500 miles). Again, estimates are difficult because there is so much variation in usage. There are owners who have never used a Supercharger, and there are those who drive long distances and use one every day. But I would say that at that price Tesla is not losing money on the whole. So I think the system as it currently exists will scale.
Personally, I think the best and most consistent route for Tesla to take would be to make all Model 3s capable of Supercharging, but charge $2000 to enable it, just as they did with the S60.