While young firms often benefit from their relationships with established firms, these relationships can be risky. Hence, during their relationship with established firms, young firms must constantly monitor signs of a failing partnership and terminate it before being in a disadvantageous position. However, discontinuing the alliance with an established firm can also be risky, especially if the young firm has limited alternative collaborative opportunities. Our study adopts the young firm's perspective and dynamically weighs the tradeoffs between the risks of continuing and discontinuing its relationship with established firms, thereby deciding on its termination. We first develop an analytical model to understand how the alliance duration (time from alliance formation to termination) between young and established firms is affected by alliance, firm, and industry characteristics. We then test the resulting hypotheses on a sample of 1,111 alliances with licensing deals formed between 159 established pharmaceutical firms and 448 young biotechnology firms during the 1986 to 2000 period, which straddles the technological discontinuity of combinatorial chemistry. Our empirical results provide partial support for the hypotheses derived from the analytical model, informing us of firms’ rational alliance duration decisions as well as their deviations from rationality. In presenting both optimal alliance duration decisions and suboptimal alliance duration practices, our mixed-method approach offers important implications for theory and practice.