It wouldn't be practical. The great advantage of the monorepo is that it encourages teams to use libraries from almost anywhere in the company. One disadvantage is that it makes splitting out those codebases almost impossible.

Has happened though; Niantic for example.

Niantic was an "internal startup", more like an X project than part of Google borg. It didn't integrate with major infrastructure like GSuite. It ran on AppEngine: https://news.ycombinator.com/item?id=12607874

I've heard numerous times about how everything inside Google is so tightly coupled to their infrastructure that making it work outside is extremely difficult.

It's true. Trying to isolate a specific product and deploy it independently if it's grown up "organically" inside Google's core infrastructure is like trying to take an animal from the rainforest and air-drop it into Antarctica with a "good luck" and a smile.

They would mean selling all documents ever printed using their product.

How did you make the connection from "Cloud Print codebase" to selling the database of all users with it?

Because why would anyone buy what basically is a CUPS implementation if not for the user data?

Sometimes companies that are doing "basically a X implementation" are still worth a lot.

That is very true. I guess I got too jaded about consumer data sharing/exchange between companies. Thanks for your feedback!