I suspect the powersat evolved as a mission justification: so, we can conceive of human-habitable structures in space, but why would we do that. From Heppenheimer's book, it seems that the colony concept emerged first, then the powersat concept.
Once you've arrived at "orbiting solar power stations built from lunar regolith", questioning your premises and realizing that the colonies really don't make all that much sense. But then again, pragmatism and overt missions never really ruled space anyway -- see Apollo and the Shuttle.
As for the technical reasons, to paraphrase Douglas Adams: Space is expensive. Really expensive. You just won't believe how vastly, hugely, mindbogglingly expensive it is.
The ISS is the most expensive single structure ever constructed, and among the most expensive single projects ever undertaken -- and that for a habitation for six people. Total cost is somewhere north of $100 billion (how far north isn't clear, the budget's spread among multiple agencies and nations).
Given that, and absent tremendous gains in efficiency and automation, the cost of solar power satellites let alone O'Neil space colonies would be in the tens of trillions of dollars, minimum. That's on the scale of the entire US GDP, if not more.
As for your observation "a lot more than actually did [happen]": Colonies in Space is one of my benchmarks for rates of technological change that didn't happen. I'd anticipated that that could well be my future, but a decade and a half past when it was supposed to be a reality, it remains science fiction. It's given me a good sense of the sorts of change which are possible and likely, and those which aren't. And no, not all progress is subject to Kurzweil's Law of Accelerating Returns.