The problem with this is that in the case of Google/FB/etc the database to be copied is so large as to make this task functionally impossible. Even if you happened to have a half a petabyte of storage available somewhere you would lack the bandwidth to exfiltrate. The problem with Parler is that it was living on a shared cloud so it was easy for someone to spin up copies and use intra-DC copy mechanisms (or even simple volume mirroring) and the data was also small enough to be within reach for anyone who was curious -- it would cost you less than $5K to copy and then export the entire Parler corpus in less than an afternoon.