Rational interactions between agents are often confounded due to disparity in their latent, intrinsic motivations. We address this problem by modelling interactions between agents with disparate intrinsic motivations in different kinds of social networks. Agents are modelled with a variegated profile over the following kinds of intrinsic motivations: power, achievement, and affiliation. These agents interact with their one-hop neighbours in the network through the game of Iterated Prisoners’ Dilemma and evolve their intrinsic profiles. A network is considered settled or stable, when each agent’s extrinsic payoff matches its intrinsic expectation. We then address how different network-level parameters affect the network stability. We observe that the distribution of intrinsic profiles in a stable network remains invariant to changes in network-level parameters over networks with the same average degree. Further, a high proportion of affiliation agents, who tend to cooperate, are required for various networks to reach a stable state.