The obesity epidemic is among the greatest public health challenges facing the modern world. Regarding dietary causes, most emphasis has been on changing patterns of fat and carbohydrate consumption. In contrast, the role of protein has largely been ignored, because (i) it typically comprises only approximately 15% of dietary energy, and (ii) protein intake has remained near constant within and across populations throughout the development of the obesity epidemic. We show that, paradoxically, these are precisely the two conditions that potentially provide protein with the leverage both to drive the obesity epidemic through its effects on food intake, and perhaps to assuage it. We formalize this hypothesis in a mathematical model. Some supporting epidemiological, experimental and animal data are presented, and predictions are made for future testing.