Due to the U.S.-China trade war, multinational firms may develop new contract manufacturers outside China to hedge against high tariffs on Chinese exports to the U.S. market. However, tariffs exhibit high uncertainty in recent years and developing a contract manufacturer incurs costs; hence, it is challenging to decide whether to develop a new contract manufacturer. We study two competing firms’ contract manufacturer development decisions in a sequential game. First, we find that multinational firms prefer to develop new contract manufacturers when tariffs are expected to rise moderately rather than sharply. This is because the value of developing a new contract manufacturer for a multinational firm is the largest when the new contract manufacturer and the existing one compete most intensely, which happens for similar tariff-inclusive costs. This implies, when taking into account competition, an overly high tariff on Chinese exports to the United States does not necessarily serve the purpose of switching suppliers from China to other regions. Furthermore, higher tariff uncertainty can decrease development value and hence the incentive to develop a contract manufacturer. When there is an upward tariff shock that induces the equilibrium where both multinational firms develop a contract manufacturer, both firms must be worse off; only when it induces the asymmetric development equilibrium is it possible for the multinational firm developing a contract manufacturer to be better off, even with tariff increase and development cost. Second, the impact of development cost on multinational firms’ incentives to develop new contract manufacturers is nonmonotone. As the development cost increases, the equilibrium can switch from both not developing new contract manufacturers to one developing a new contract manufacturer. Third, when the future tariff is expected to be high, multinational firms should diversify their development strategies in environments with fierce competition. Furthermore, increasing competition is not necessarily harmful for multinational firms. We also extend the analysis to consider asymmetric development costs and unobservable wholesale prices, and find the major insights are robust.