It’s designed with client-sided prediction and lag compensation, so this shouldn’t be too much of an issue.
So?