What's better-- the right project over budget, or a broken nothing that is under budget?
That right project over budget is clearly what we want! I do claim that teams who have don't have a way to iteratively validate their increment are just going to deliver the wrong thing over budget though.