I think this is one of two fundamental flaws in your reasoning. First, yes, R&D is incredibly, incredibly expensive. I don't think anyone doubts that, and I find it just hysterical that x or y "social startup" can burn through far more money than that and never have to answer questions about the value of their R&D.
We know what Tesla is doing, and we know that it's difficult. It should be no surprise that it's expensive as hell. This isn't to say I'm betting on their success, it's to say that "R&D is expensive" isn't an argument against Tesla or its potential for success.
Quite the opposite, it seems their model has done a decent job of offloading a lot of R&D costs by, as I said, selling automobiles on the side.
Now, the second flaw ("nor can he raise the price of gas") isn't that the statement is false. He can't. But the laws of thermodynamics paired with the laws of economics can, and will, forever. Further, when it does happen, the success of his cars (or "toys of the idle rich") is irrelevant: he owns the infrastructure and he owns the IP necessary for everyone else to leverage the inevitable growth in fuel prices.
Also, to address the battery capacity loss argument:
http://www.pluginamerica.org/surveys/batteries/tesla-roadste...
Independent study on the Roadsters.
126 vehicles (of 2500 produced) reported a total 3,198,749 miles traveled. Tesla's projections said the battery would have 100,000 miles driving range, so at 50,000 miles the projection was to have 70% capacity. In reality, an average Roadster has 80-85% original capacity after 100,000 miles. That said, there was considerable variation between batteries' health, so take that as you will.