I can think of a way to help with the problem, but I don’t know how hard it would be to implement.
Create some sort of trust score, where instance owners rate other instances they federate with.
Then the score gets shared in the network. Like some sort of federated whitelisting.
You would have to be prudent a first, but not do the whole task yourself.
You could even add an “adventurousness” slider, to widen or restrict the network based on this score.
Meta will face daily fines of €89,500 if it doesn’t comply with the order.
Bet they can write it off as expenses.