While proof of stake does have a lot of advantages over PoW, it cannot handle more than 1/3 of the voiting power acting incorrectly. This is a well known hard bound on BFT algorithms.

In terms of implementing PoS, most BFT algos have traditionally scaled poorly (HotStuff which is linear in communication costs relative to the number of nodes/validators is very recent). In addition, going offline for validators, even accidentally, hurts the network and slashing PoS schemes often have uptime requirements. Validators also need to have the infra to be hidden and not DDOS'd.

tl;dr, Proof of stake is still decently centralized, but a lot of work is being done to scale it.