Okay so I've dug into this further and figured out what causes this to happen. It would be good if some other users can test similarly to see if this is happening for everyone across the board.
The issue has to do with the new smart tempo feature, and namely that smart tempo now works with multitrack recordings in 10.4.2. I have never found a need to use that feature set. Therefore, here are how my smart tempo preferences look. My understanding is that these settings should basically turn the feature "off".

After I noticed this normalization issue, I also took notice that there are a bunch of files piling up in my audio files folder called "Smart Tempo Multitrack Set x.aif" .

What seems to be happening is that any time you record two or more tracks into logic, it creates one of these files which I can only assume serves as an analysis file for the smart tempo feature.
Dragging and dropping this into logic next to 2 recorded files shows that this file is a summed / normalized version of the tracks just recorded.

Again, assumptions. But seeing as the recorded files don't change from their recorded level, the status bar that myself and other users are reporting showing normalizing isn't in fact normalizing the recorded tracks, but instead normalizing this analysis track. Further proof of this can be shown by inspecting your undo cue just after a multitrack recording.

And for those who don't believe that this is happening:

Mind you, this only seems to occur when the process of generating this analysis file is a resource heavy one. For some users like myself who record long multitrack audio at high sample rates, this is literally every time and it considerably slows down performance, is disruptive for workflow, and essentially doubles the size of my projects.
Suggestion for Apple, provide a way to FULLY turn this feature off. If the user has no intent to use the feature, then please remove the resource-heavy background processes and files piling up in the background.
Again, would appreciate if other users can test and see if the same thing is happening.