1B search queries are made every day using Google. If the average length of a Google query is 10 characters, that's 10GB of data logged only for the query itself. I'm assuming they optimize the way the store it to avoid repetition in other information about the query like information about the user, browser, etc. and compress it but still. Now factor in other data like visiting websites, watching youtube videos, e-mails, uploading files, downloading files, logging in, logging out and probably a bunch of other stuff. And according to this sub, they store it all and they store it forever.

I seriously don't understand why would they ever need to do this. It seems like it requires enormous amounts of storage and compute for honestly questionable benefits. "They have to due to the law" Ok. Not forever though? "They want to sell you stuff" Surely there are better ways to build up a recommendation model that to keep every single thing a user has ever done since their first time interacting with the software and then process it whenever they do anything, even with the best algos in the world. "It's not a big deal for Google they can manage this amount of data" Yes but why?

The only thing I managed to came up with is that they sell it to whoever wants to sell you stuff too but then again, every single detail and the revenue from that is worth all this hustle?