Nuclear proliferation has in part led to a vastly-different security environment from that which existed during the Cold War. While Schelling's theory of deterrence through mutually-assured destruction established a precarious stability for forty years among the great powers, its assumptions render it of limited utility for evaluating the effects of nuclear proliferation below this level. Using an elaboration on Powell's bargaining models of conflict, this paper shows that the introduction of nuclear weapons in second-tier states radically alters the future distribution of power. These states therefore face a commitment problem vis--vis conventionally-stronger neighbors, inviting preemptive attacks on soon-to-be nuclear powers. This paper analyzes the dynamics behind this problem, as well as offers a theory of great-power intervention that explains the empirical lack of preemption.