Kant is clear that the concept of the ‘highest good’ involves both a demand, that we follow the moral law, as well as a promise, that happiness will be the outcome of being moral. The latter element of the highest good has troubled commentators, who tend to find it metaphysically extravagant, involving, as it does, belief in God and an afterlife. Furthermore, it seems to threaten the moral purity that Kant demands: that we obey the moral law for its own sake, not out of interest in the consequences. Those commentators brave enough to tackle the issue look to the concept of the highest good either to add content to the moral law (Silber), or to provide rational motivation, in a way that does not violate moral purity (Beiser and Wood). I argue that such interpretations, although they may be plausible reconstructions, are unable to account for certain conceptual and textual problems. By placing Kant's thought against the background of medieval theology, I argue that the hope for the summum bonum is irreducibly important for Kant, even where its function is not that of providing the content or motivational force of the moral law. Kant is not only concerned with the shape of our duties and motivations, but the shape of the universe within which these emerge.