Im not really well versed in energy generation - do you have an estimate how many solar panels would be needed to power one supercharger at full capacity?

A couple acres? No joke. To generate the 50-80,000watts necessary for each stall you will need city-blocks worth of solar. If it's cloudy or at dawn/dusk, more. If you want to charge cars during the day and night, triple the number of panels and get some massive batteries.

Say 120 watts per m2. 80,000/120 =~ 650m2. That's the ideal. Then for dawn/dusk, 1200m2 and triple that to run it at night ... 3600m2 ... just shy of the 4050m2 in an acre. But even that is 3600m2 perpendicular to the sun not on the flat ground. To get that 3600m2 in the real world (not on the equator) you will have to spread the panels out on a much larger lot (much depends on the location and time of year). If I were buying land, I'd estimate 2-2.5 acres for solar per stall.

This is ridiculous. People aren't charging 100% of the time, and even when in use, the chargers burst to 80kW, but they don't average 80kW. All you need are some batteries to level the load. Assuming 20kW average, and stall being used 20% of the time, that brings your calculation down to 4,000/120 =~ 33m2. Accounting for night and bad weather, let's quadruple that to 133m2. That's about a quarter of an acre for EIGHT stalls.

I skipped many other multipliers that negate any error in the needed wattage per stall. Top of the list: latitude. Demand will be constant throughout the year, but solar output can be radically different depending on the season. Then, is the horizon flat? If it isn't, then you aren't getting as many hours of sun. Then whether the panels are fixed or sun-tracking (tracking is better but require more space/maintenance). Tracking rigs are also much less space efficient when the sun is low. They end up shadowing each other for much of the day. Access roads, setback from obstructions (treelines) panel output degrading over time, energy losses in/out of batteries ... there are scores factors to undermine the per-acre efficiency of small solar farms located near to consumers. Imho tesla would do far better by building large solar farms in perfect locations and hooking them into the public grid. Panels at the charging stations can be the window dressing.

Just a clarification-- the chargers burst to that 80kW range but they don't maintain it. Typical charging is closer to the 10-20kW range.

In addition people are not constantly charging.

Budgeting solar is like chopping firewood for winter. You cut as much as you think you need. Do that again. Then the whole process again.

There are so many factors that can diminish the amount of power available at a particular location that the only reasonable thing to do is overbuild. Imagine a line of teslas at an upstate New York charging station. They cannot get to work without power. But it's raining. And it's winter. And it's 6am. And that treeline has grown a foot higher, pushing dawn back an hour. You cannot have too many panels.

Lots of them. A few hundreds, I think. Viability may depend on site. But if there's space nearby for a reasonable array, there doesn't seem to be any reason not to use it no matter how small it gets (any size will result in savings), unless your electricity is dirt-cheap, or the place is so infrequently used that for a significant portion of the time, nobody is charging (which would waste a portion of the local generation and worsen the ROI).