> Is it correct that Reddit hiked its API prices in response to realising their data is worth a fortune (since it can be used to train/update LLMs)? Is this understanding correct?
This was one of their stated reasons, but as you note, this rationale doesn't make a whole lot of sense considering that most AI datasets would just train based off scraped Reddit data, which wouldn't require use of their official APIs.
In practice, it was likely to kill third party apps, in a copycat move of what Twitter did earlier this year.