If the top 10 institutions have the vast majority of the top 1% of grad students, I'd expect they produce the vast majority of professors.
If you're a top student who chose a slightly suboptimal grad school for personal reasons (to be near a significant other, taking care of a family member, already have a house somewhere, etc.), you may not realize that the game became cripplingly hard for you until you're near the end of your program.
The top universities often: (1) send grad students to conferences (including travel costs) whether or not they have results to present that year, (2) hire lab techs and admins, which allows grad students more time to do actual research, and (3) make it more likely that the student gets a grant proposal accepted by the NIH/NSF/other agency.
And you know, maybe you hold out hope because your advisor let you and your lab mates volunteer stuffing all the swag bags the night before a conference that happens to be in your city that year so you can go talk to all the other folks at the top of your field. Maybe you actually get a grant funded, unlike your classmates. Maybe you put in that many more hours so you can handle your share of the grunt work on top of research you'll actually get published. No weekends for five years. Do you do it? Maybe.
Furthermore, articles would be greatly puffed up if they had to preempt random arguments which uncritically mention notions like "top 1% of grad students". (Kind of depresses me that I'm immersed in such a tech/managerialist culture that this argument popped into my head too while reading the article, requiring me to spend time evaluating it.)