As I understand it, Google does a decent amount of rendering of a page before indexing; this a) allows it to index content loaded by JS and b) prevents some way

As I understand it, Google does a decent amount of rendering of a page before indexing; this a) allows it to index content loaded by JS and b) prevents some ways spammers show Google different content from users. Perhaps Google's main way of storing a page no longer matches something that can be easily served as a cache page. This might be a way to remove a legacy copy of each page and reduce storage costs.