How would that make sense? (And do you mean "first x megawatt-hour produced or what?) Unlike coal or nuclear, solar has almost only upfront costs (production and installation), so it stands to reason that you don't reduce the rate for already installed units. (Though inflation and cell degradation will reduce real annual income over those 20 years.)

The calculation in the blog post you link is funny: it doesn't even attempt to substract a market price from the feed-in rate, declaring all of it a "subsidy" (which, then again, is standard practice in propaganda from the German energy giants). (It would not be an easy calculation: you'd need to project future inflation, cell degradation, consider the diurnal and seasonal change of both PV output and market prices; and correct for grid use, which is zero for own consumption.) Whether even the thus the overestimated sum from that blog post is equivalent to building 6 EPRs, was already questioned by Jérôme; let me add operating, fuel and dismantling costs.

_{*Lunatic*, n.
One whose delusions are out of fashion.}