The Supercharger network, with 270 sites in the US today, and due to be doubled by the end of 2017, is already beyond the approximately 200 locations that Tesla Motors originally forecast as being necessary to enable long distance travel on our highways. It would be unwise to presume that having perhaps 150% saturation beyond that 200 locations by the end of 2017 is intended to be solely for use by Model S and Model X. No. That expansion is specifically to allow use by Model ☰ buyers. So, yes, by buying today, you are paying for the infrastructure that will be used by future buyers. That also includes Tesla Stores/Galleries, and Service Centers. It also includes, I hope, what would be Tesla Waypoint or Tesla Depot locations that catered to long distance travel by way of electric vehicles, some day.
When I buy an ICE, there is an NPV associated with my energy consumption. Business of all sorts make decisions not only on what customers will pay up front, but the demand that will exist in the future. Would you also be saying that I'm creating economic value in infrastructure for gasoline stations when I buy an ICE based vehicle?
If so, then OK - you're right. Not only do I benefit, but so do others. This analogy could be extended by my need for groceries... Wal*Mart does planning based on demand, so maybe my purchase of apples is supporting apple growing/distribution infrastructure, which future buyers of apples will benefit from, and ultimately flows into the creation of Wal*Mart stores.
Tesla, assuming they have access to the capital markets (which they do), can choose to either finance the infrastructure's initial build out by asking for money up front or just know that the demand will be there (and they get revenue by charging ongoing fees irregardless of how - pay per use, per year, whatever) and utilize the capital markets to fund the initial build out and ongoing use of capital. At the end of the day, it's a wash if you believe that the markets are efficient allocators of capital and that it's probably more efficient do it by the billions rather than by $2,500 at a time.
Given that they have easier access to the capital markets (look at the last equity rise) than in the past, I think Tesla is likely to charge on an ongoing basis rather than build in an infrastructure development cost into the price. At least, it'll be an option. They are probably going to take their expected per kwh cost, add in the expected economic depreciation of the equipment on a per charge (including non-used time for weather impacts, etc.), and ultimately charge that to the customer. Any variation will be based on expected usage calculations (how many charges will the average user consume, etc.) or even being sold below cost to encourage the SC network use.
I know I've shifted topics, but I think the financial aspect from understanding the financial statements and their access to the capital markets is critical. Things are very different than they were when they were just starting to build out the SC network, and so maybe they did the calculations posted earlier in this thread at one point, and today they'll still be similar, but applied in a different manner.
Today, they shouldn't be viewed as splitting up the payment into infrastructure and electricity - it should be viewed as just as "all-in" cost on a per charge basis, which they will likely display and sell on either an actual underlying usage basis or time based subscription, or some kind of similar variant. If they do a 1-time up front fee, it's going to be purely based on this "all in" cost and an estimate of the number of charges - not specifically targeting a certain amount of cash flow for infrastructure build, because they have easier access to the capital from the markets today.