Well, it might be extraordinary if it were actually a fact, but as far as I can tell it isn’t.
Using data from the University of Virginia Historical Census Browser, and looking at the 1790 Census (when the process of Northern abolition had only just gotten going and the thesis therefore has the best chance of being correct): The indisputably Southern states of North and South Carolina had a total population of 644,078 people, and 135,218 of those people were members of white slaveholding familes (just under 21% of the population). In the Northern states of Connecticut, New Hampshire, New York, Pennsylvania, and Rhode Island, there were 1,222,518 people, of whom 73,959 were members of white slaveholding families (a little over 6%). By both absolute numbers and per capita, there were more slaveholders in the South than in the North.
If you insist on counting Maryland as a “Northern” state–which nobody in the antebellum era would have, but what the hell–then the “North” now has 1,542,246 people, of whom 145,127 were members of white slaveholding families, or 9.41% per capita white slaveholders. Now the absolute number of slaveholders in the “North” is higher, but the per capita of slaveholders in the “North” is still drastically lower than in the Carolinas.
(Those are all using total population, which includes slaves, and of course slaves didn’t own other slaves. The North had a free population of 1,193,855, of which members of white slaveholding families made up 6.19%, the South had a free population of 436,201, of which members of white slaveholding familes made up 31%. If you ahistorically insist on counting Maryland as a Northern state, the “North” now has a free population of 1,410,547 people of whom 10.29% were members of white slaveholding families.)
But wait! Even for 1790 you’ll note I’m leaving out several states. That’s because there are major gaps in the data: no information is available on members of slaveholding families in half the states and territories at the time. (I dunno, maybe rodents got at some of the old census records or something). So, I’m leaving out New Jersey, an indisputably Northern state which was slow to abolish slavery and still had slavery well into the 19th century.
But I’m also leaving out Virginia (and Georgia), neither of which has data available on total members of white slaveholding families. Virginia was easily the most populous state at the time; and since in 1790 almost 40% of the population of Virginia were slaves, I’m going to go out on a limb and say there were probably a hell of a lot of members of white slaveholding families in Virginia.
Also, Massachusetts is another state with “N/A” data on members of white slaveholding families; but in fact by 1790 slavery had already been abolished in Massachusetts by “activist judges” and the 1790 census thus shows a slave population in Massachusetts of zero. Massachusetts had a total population of 378,556 (and also a free population of 378,556, natch), which puts the per capita numbers of members of white slaveholding families in the North at under 5%.
So, the South therefore clearly had more slaveholders than the North in absolute numbers and per capita.
Pretty much any way you cook the data, slavery was concentrated in the South–and that only became more and more true as the process of Northern abolition continued into the 19th century.