I think you're mistaken in saying "If we are taking about Great Migration immigrants to New England and the first few generations, then yes the majority are already in wikitree."
What happens is people start "doing genealogy" by finding research that's already been done and connecting to it. Things have already gone awry right there, since the "connecting to" stage often involves a lot of hopeful assumptions. But once the connection is made, people "do research" by looking up more family trees. Since they start out from lines that have already been researched, they find that - surprise! - "their" line has already been researched. Meanwhile, their actual ancestors languish in obscurity. It's not that the majority of American colonial ancestors are in wikitree, but rather that the majority of late 19th and 20th century American genealogies have been copied into wikitree.
That should be obvious just from simple demographics. Take the name "John Smith". Smith has a frequency of around 1.5% in English contexts, and John was about 20% in the relevant period. There were around 250,000 people in English America in 1700, let's say half male, so there should be something like 400 John Smiths in America in 1700. I see about 10 John Smiths total on Wikitree who could have been alive in 1700, and probably only half of those were in America (hard to tell with some because of the sketchy documentation). Let's say we don't care about people who didn't leave descendants and consider that wikitree is mostly concerned with those who did. If we say only about 1/4 of those left descendants (a fairly typical figure), then Wikitree should have around 100 John Smiths in America, but it actually only has 5.
I agree, wikitree represents famous and well researched "gateway" ancestors well, but it's not even at 10% of American colonial ancestors ca. 1700. Even if my estimates are off by a lot, Wikitree is obviously an order of magnitude away from having the density of coverage that you think it does.