Another speculation is that it's about preventing snowballing. The idea is that if you have a tiny number of planets, your "demand" for leaders is below the limit and you can afford to staff each planet with a leader, maybe 2 for assist research.
The problem is that "arbitrary limits" NEVER prevent snowballing. They only slow it. And I don't know about you, but I'd rather have actual catch-up mechanics that reward you for being "behind" than "A slower death".
I think this is the wrong framing. For a very long time now, Stellaris has made no meaningful attempt to reward players who confine themselves to a small number of total colonies. Yes, there's empire size from colonies/systems, but that's a relatively trivial expense for most empires, and I suspect deliberately so.
Instead, what's encouraged is herding of pops. The big emphasis has been on having a small number of *core* colonies. These are the colonies where most of your pops live, where most of the resources are produced (as far as districts and building slots allow), and where you invest heavily in infrastructure, whether that's Ecumenopolis, orbital ring or planetary ascension. You can have other colonies but they are mainly pop farms, or defensive outposts to buffer your core, or surplus colonies that are left to subjects; they will never be efficient at most resource production, so you try not to invest resources in them that are scarce at an empire level. Basically, if we ignore vassal swarms for the moment, empires end up looking more and more like Australia, a continent where half the population lives in three cities, and even among the other half, most still live in cities; farmers and miners are an important part of the economy but a small minority of the actual workforce (for instance the country produces enough food for around 75 million people, but only has about 150,000 farmers).
The most important thing though is that you build up your core colonies to have a lot of pops; since you won't get those pops in a timely manner by "going tall" and staying inside your core area, you're encouraged to go out and harvest population from a larger number of colonies. This is how the game pushes you into "wide" play. (The difference with Australia here is that in Stellaris, you can't get large amounts of immigration from outside your territory, and the fringe colonies will collectively spawn a lot more pops than "space Melbourne" because the game's model of population growth is much more based on number of colonies than reproduction of an existing population.)
Leader cap is just another plank added to this model. The devs have effectively declared a design intention that most colonies will have no governor. You choose a few planets to be "core" and have a governor (and/or scientist in orbit, who is basically a special type of assistant governor), and maybe you try to make those choices in a way to maximize a kind of hinterland bonus from governors in the sector.
Where the model falls down I think is with subjects, because there's no cap on the number of subjects you can have (at least, no cap that can't be eliminated at the cost of 1AP), and all the factors that favour a few core colonies in your own empire, also favour having many small vassals over a few large ones (plus some additional factors on top, like trade agreements, holding limit, and per-subject bonuses from Scholariums and Bulwarks). Governors are effectively added to this: every micro-vassal planet can have a governor, no problem, even if your vassals collectively own hundreds of colonies, because they don't use *your* leader cap. Vassal swarms are effectively anti-"meaningful choice", and the more the rest of the game tries to go in a "quality over quantity" direction, the more vassal swarms will distort and undermine this idea.
