Would you sac a knight to prevent castling?


King safety tends to be a dynamic (think short term) advantage.
Material is a static (long term) advantage.
What makes chess so interesting is trying to weight one advantage against another when they're so different. In some positions what you describe is very good, and in other positions it's very bad.
During the first 10 moves it's unlikely you'll have a large enough lead in activity to justify sacrificing a knight like that, but it's definitely possible. What matters is that you have a followup that can keep generating threats until you either win back the material or checkmate them.