> This is true for a single panel. But the amount of sunlight which hits an acre of land is constant. If the land is 100% covered with panels, the panels will collect 100% of the available sunlight.

That is not true from my understanding. The increased solar angle of incidence effects how much of the energy reflects back into the sky. Having more panels next to it without a gap won't change that. Yes you can fit more panels in the same area without putting them on an angle but they will be quite a bit less efficient because more light will be reflecting back up per sqft of panel which is what matters cost wise.

There is a small loss of efficiency (<2%) at extreme angles, but it is not too significant[1]

[1]: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6611928/#:~:tex....

The anti reflective layers and texture on the panels are usually pretty good, such that then reflection is less than the area loss due to the angle.

Yes, one of the many areas of improvement of PV modules in the last couple of years has been reducing reflections from the surface of the glass and from the glass-cell boundary in panels. Also changed cell technology (n-TOPCon) helps with reflections and recombination within the cells.

it's true that you get 100% of the sunlight that falls on the land, or near enough (albedo is not literally zero at any angle) but you need more solar panels per output watt

it'll be interesting to know if they're derating the nameplate capacity by cos(latitude) as they should be

Their agreements with their local grid operator will specify power delivery. Marketing numbers like an individual panel's nameplate output are not important at utility scale, which this is designed for.

probably the local transmission companies are interested in not only average ouptut but peak output, but also, there were a number of numbers in the article cited as "per watt" which look like they might be talking about peak watts