To me, the reason why the game industry has little open source libraries is because games are usually a once-and-done approach to software. So open sourcing part of your game has little benefits because you're not going to go back update the open source libraries. Whereas other software is a constantly updating behemoth, so open sourcing your software can get small beneficial increases from external changes to your open source code.
The profit motive doesn't really seem like the main reason, since most open source software makes no money. Facebook doesn't get money from open sourcing React for instance, but they do get functional improvements to React that comes from it being open source and having more people help work on it.
> games are usually a once-and-done approach to software
This is just false.
The most obvious example of this is live service games - these have been commonplace for over a decade, and have existed for 20 years (WoW in 2004 in my mind is the turning point for that). Slightly less obvious are PC multiplyer games that predate this - as an example Counter Strike was released in 2000 and was "patched" a few times before steam existed.
Even going way back, one shot single player games have been built iteratively since the 90s at least, if not earlier. The idea of a game "engine" is reusing code between games - The LucasFilm games all used similar engines to speed up development from the late 80s. Even today, modern games like Call of Duty have their roots in Quake's game code from the late 90s.
The profit motive doesn't really seem like the main reason, since most open source software makes no money. Facebook doesn't get money from open sourcing React for instance, but they do get functional improvements to React that comes from it being open source and having more people help work on it.