A big reason was that the instance was always going down due to ddos attacks, shitty federation code. Kapow got tired of it. It's more decentralized here so easier to keep up.