Cost-over runs aren't relevant. Cost is. Case in point, American Muni Power is refinancing its costs for the wildly jacked up Prairie State pulverized coal plant, this week. Including the mine, S&P is remarking its 1,582MW are up to 5 billion. That's $3.16/watt, before sub-100% capacity factors. It isn't relevant who broke their estimate, or that this plant was supposed to deliver @~$1.50/watt. It's where $$ landed. On that score, Vogtle has already had added costs, and sits currently at about $7 ($7.70 @90%CF). Topaz solar, perhaps the biggest CA commercial PV, is looking at $2bb, for 550MW, but its solar so you have to wash in the much lower capacity factors associated with sun (15%-25%). That takes $3.63/watt up to $14.50 per watt of usable solar capacity. This also carried a price tag for a couple thousand acres of, presumably cheap bright, desert land.
You want to point at cost over runs, but I say look at cost.
Your previous points are well taken, with roof tops and car ports. I met some of Walmart's team, a few days ago. >100MW on their stores, but no plans for car ports because of the price and liability, versus roof mounts and limited ground mounts. Maybe they will get there, but consider just Topaz (try Google image) will require about 8 million panels. This is what raises my blood temps, when other forms of CO2-free energy are targeted as a horse trade. The perspective is impractical.
FYI, Those images of topaz replace a little more than 1/10th the output of the coal plant above (80%CF, on 1,582MW vs. 25%CF on 550MW STC-rated). We can do this math over, and over, again to see the wood solar has to chop, or clarify what we ask of ourselves if we want to go after the fossil elephant.