Assume you've got a couple dozen slots, charging 100 cars a day. (Far, far fewer than an equivalently-sized gas station. If we're going to transition to an electric-only transportation infrastructure, then car changing stations are going to become a common sight)
That's 9 megawatt-hours a day. A random solar panel calculator I found guesses that would need ~2000KW of solar panel capacity. So you'll need 8,000 250 watt panels. That'll cover 9,600 square metres, (2.6 acres-- if you're lying them flat, which you won't. A big dedicated solar generation station will use tilted one-axis tracking, which will use a lot more ground space) and cost you (at $410 a panel) a cool $2.4 million.
(Google's big solar installation at Mountain View, the one that got so much press, was only 1700KW. Our tiny electric car charging station has to be 17% bigger than it.)
Electric car charging uses a lot of power. Solar is great, I love solar, but it's not space-efficient, or cheap. Either Tesla is going to build a national solar generation network on the scale of Germany's move into solar power, all on their own dime, or they're lying through their teeth about charging draw/solar production breakeven.
I don't think Elon Musk would green-light the "100% solar power" line without doing the math. Again: you can argue about whether he's lying, but I can guarantee you he knew exactly the scale of what he was promising. That makes it all the more impressive, imo.
Re:lying, I'm imagining at some point they'll transition over to "we buy 100% solar power for SuperChargers", which is fine by me since it's functionally equivalent to what they're doing now.
Again, this is a guess based on a CG rendering on their website. That drawing doesn't constrain Tesla's engineering talent in any way – even without using extra land they could put up solar shade structures over adjacent parking spaces.