A large-scale electric vehicles (EV) penetration will consume significantly more electricity, which may overload power distribution systems and charging stations (CSs). We propose a multiagent optimization approach to study the multistage decentralized distributed generation (DG) and CS planning process considering the decision makings of the DG investors, CS aggregator, and distribution system operator (DSO). Furthermore, we develop an exact convex reformulation of the proposed multiagent optimization problem to significantly improve the computational efficiency. Through numerical examples, we found that with the increase of EV adoption, CS investors expand the CS capacities significantly towards the later years of the planning horizon, whereas the DG owners expand their generation capacities in response to both load and charging demand (CD). The model is able to describe the close coupling between stakeholders and efficiently identify the equilibrium investment patterns of DGs and CSs over space and time to satisfy the growing EV, CD, and other existing power loads. In addition, the modeling framework can allow DSO and government agencies to evaluate viable system evolvement paths and better understand the system needs of energy generation and charging infrastructure in order to achieve a higher level of EV penetration.