Both the entry and withdrawal of partners have received considerable attention in the venture capital (VC) literature. However, they have been viewed as more or less isolated dynamic choices. This study filled this research gap by analyzing the relationship between new partner addition and incumbent partner withdrawal in VC syndicates. Although the literature on VC partner selection suggests that adding partners can improve the performance of VC syndicates by emphasizing the expected contribution of VCs, the expected development will not occur or may even be reversed if the VC syndicate also experiences the withdrawal of the incumbent partner. We proposed that the VC withdrawal decision is a risk-benefit trade-off. The syndicate’s portfolio similarity with new partners promotes competition between them. This risk of knowledge leakage through the relationship with the new partner jeopardizes the incumbent partner’s competitive advantage. In the meantime, the syndicate’s portfolio similarity with new partners measures the extent to which the knowledge contributed by the new partner matches the needs of the syndicate and can be absorbed by the syndicate, which will enhance the common benefits.

Thus, there is a U-shaped relationship between the syndicate’s portfolio similarity with new partners and its incumbent partners’ withdrawal possibility. Furthermore, we considered the negative moderating role of partner network characteristics, namely, the relational and structural chasm. Our research not only suggested the dependence of network dynamics to complete the theory of network evolution but also provided advice on partner selection and syndicate governance. We tested our hypotheses using the dataset of VC investment in the United States between 1985 and 2016, based on Thomson Reuter’s VentureXpert Database and ORBIS Database.