The link you provided does not in fact back up that claim. It shows a demand for 15.1m new homes, with 8.4m of those being one or two person. That’s barely over half - nowhere near 80%.

In fact it does. The projected 15.1M new homes needed is not the same number as the projected number of new households, for reasons the text explains clearly.

It projects 12.1M new households, 8.4M of them either one-person households or married couples with no children (which presumably means a two-person household, although I suppose it's possible some of those might have a third person living with them, e.g., an elderly parent). Only 1.6M of new households (married with children) are obviously more than two persons. The remaining households are 0.6M unmarried with children, which could be two or more persons, and 1.5M unmarried living with other adults, which could also mean two or more persons. It's not impossible to get somewhere close to 80% from that.

At any rate, the comment I responded to asserted that the expert in this field said something obviously untrue, yet offered no evidence except his own personal experience shopping for a suburban home in one particular American city, as if that proves anything at all.

"This source supports the claim that "something like 80%" of newly formed households will consist of only one or two persons, at least for the period 2018 through 2028."

It kinda makes sense that newly formed households will consist of one or two people. Does it mention how long they are expected to stay that way?