The one state with a truly anomalous result--NC, where Trump got about 625k more votes than the "Black Nazi" for Gov, is easily explained by the uniquely toxic nature of that candidate, with ~400k of those Trump voters picking the Democratic candidate, and the rest leaving that office blank.

2012 2016 2020 2024
AZ 39k 163k 24k 155k
MI 345k -- 8k 80k
NV 5k 18k -- 64k
PA 111k 77k -- 145k
WI 29k -73k -- 38k

So looking at these results, none of them suggest any really anomalous overvotes for Trump.

B) Ignoring the historical context of previous elections. Taking my projected overvotes for the five states you looked at in this election and comparing them with the three previous elections, we have the following table of Republican Presidential overvote results (x1000):

Where your and Spoonamore's analysis is fundamentally flawed:
A) Assuming all the votes have been counted in this election--they have not, particularly for the various Senate races. While you've got the PA overvote about right, you've overestimated NV by 10k, WI by 16k, AZ by 20k, and MI by 37k.

While you're correct I should have provided a more detailed breakout of Republican presidential overvotes (which I'll do below), my initial point was simply to demonstrate that presidential overvotes are not at all uncommon in the historical context.