We study capacity planning in the context of Infrastructure as a Service (IaaS) cloud services. Concretely, we propose models for optimal pricing and virtual machine allocation to maximize revenue while maintaining service reliability. Our models are based on current pricing schemes used by predominant cloud services providers in the market. These providers frequently face slack capacity and low utilization in resources, so that they require better planning in the allocation of non‐revenue‐earning resources that could significantly affect profitability. In particular, we propose two novel hybrid service models, each combining two basic types of cloud services: reserved‐instance (contract) and on‐demand, and provide conditions under which they become dominant. We compare analytical properties of different service models, derive closed‐form solutions and structural results to assess the strategic impact of the different models, and conduct numerical analysis to test how optimal solutions change with respect to critical system parameters such as service rate, market size, and service reliability.