An argument might be that they played to the level of competition. When lights shined brightest, they excelled because they knew the competition has harsher. When they played easier teams they used it as opportunities to try rotations and schematics out. Which could be why they destroyed the better teams they played, but as you point out squeaked by teams like Minnesota, Penn State, and Oregon.The recent rise in metrics is more due to coaching actively working to engineer the analytics scores now than it is NIL.
NIL has allowed teams to buy some success, but has also resulted in a lot of teams spending tons of money and getting not a lot out of it (eg, Kentucky). We know from history of massive spend 'super teams' that they tend to flop just as often as they succeed, it seems.
But, like... is Duke's success due to NIL? They were pretty good before...
I wouldn't even call Michigan last year a 'super team'... lost to an unranked Wisconsin team, squeaked by a bunch of horrible-to-okay teams like Wake Forest, TCU, Penn State, Oregon... even Northwestern and Minnesota tested them. Nearly lost to Iowa and Ohio State. Nearly lost to that same Wisconsin team again.
I think there was even a metric passed around that displayed if you played better/worse against top level teams or bottom level teams, and Michigan trended to struggle against easier teams, but showed strength against the tougher teams. Illinois on the other hand I think was neutral on harder teams but was very strong on easy teams. As demonstrated in our late season skid before the tournament. Now I think factors were at play for us, like injuries, etc. but dominating the elite teams and sometimes struggling but still winning against lesser ones can get you in trouble by messing with fire at the wrong time (like in March), or can get you places like the winners podium.