Business relationships are not fairy tales. Even though some business relationships last for years, most of them are not a lifetime commitment: “One day you do business together and everything is fine. The day after, it’s over and you don’t even give a damn about the other guys” (interview, businessman, 2007). The constant possibility of breaking a commitment is not an abstract moral code that would encourage market professionals (Cochoy, Dubuisson, 2000) to seek their self-interest with guile in any circumstances. Business relationships are indeed fragile, but their fragility is rather a social fact that most contemporary economic actors accept with no difficulty (Chantelat, 2002; Surubaru, 2014). However, this social fact cannot be reduced to individual opportunism. The main aim of this paper is to understand how this social fact is taken for granted, especially by actors who officially consider themselves as being involved in “strong industrial partnerships”. Unlike highly competitive markets, industrial partnerships rely on physical and human investments that are specialized and unique to a function. In this case, breaking the commitment seems at first sight very difficult, because there is no other industrial partner available on the market. However, this asset specificity does not necessarily “keep the partners tied” or prevent them from opportunism. On the contrary, industrial partnership are very complex economic situations, where exit, voice and loyalty are always mixed together (Hirschman, 1995). These complex situations provide interesting insights into how economic organizations actually valuate their partners (what do they think about theirs business partners’ behavior, how do they think about the future, what subjects are important to them during the contract’ negotiation, what are the main problems that occurred during their last exchanges, how do they manage to defend their interests when partners do not respect their obligations, etc.). Despite the importance of this type of issue, very few sociological studies provide a satisfactory explanation of it. In this communication, I first show why dominant sociological approaches fail to provide an appropriate analytical framework. Then, I explain how a Weberian approach draws a more realistic picture of industrial partnerships. My empirical data from the French defense industry provides a highly relevant case study for this purpose.< Réduire