Yes, which is why the least skilled developers maintain the products, which is why they eventually get canned. This is not inevitable in software —- it is specific to Google. At business software companies, for example, maintenance is not career suicide, so you have very senior engineers who work on improving the engineering systems to detect and fix bugs, leading to a more effective long term software maintenance capability.
Maintaining and improving core systems is what entire large teams in Google's Ads, Search, Core, etc. teams do, and yes, they get promotion for it. It's often not very sexy work, but it is recognized.
It's really the consumer-facing & consumer hardware teams that have this novelty-seeking-promotion problem, IMHO. (I've worked on both)
Promotion rates in Core Dev are higher than normal across the company. The idea that you cannot get promoted writing tooling or making engineers efficient or keeping the codebase clean is simply not supported by data.
What it sounds like is that if you're maintaining things it needs to be something important to the business to get promoted, but if you create something new then it doesn't have to be.
I'd edit that to say that it is easier to incorrectly believe that something new is important to the business but that you really do need to make the same impact argument in both cases.
You either build some shit or end up not getting a promotion / new job.