States reap rewards when sharing intelligence, including valuable insights on threats that otherwise may be impossible to obtain. States also incur costs when sharing intelligence, namely the potential defection of a state sharing received intelligence with an unauthorized third-party. Therefore, states invoke a cost-benefit analysis on sharing intelligence to determine the potential for an intelligence alliance. These costs and benefits are magnified in multinational intelligence sharing alliances, thereby presenting scholars with an interesting puzzle: With the heightened risk of defection, why do multinational intelligence sharing alliances expand or restrict membership? I develop a theoretical framework to understand why a multinational intelligence sharing alliance expands or restricts membership, and I argue that admission requires a shared threat, high degree of trust, and complementary capabilities. I test this theory through comparative case study analyses on Australia, India, and Denmark in the context of the Five Eyes alliance from 1940-1960. Ultimately, I find support for my novel theoretical framework, which serves as a promising model for future scholarship on multinational intelligence sharing alliances.