In this study, we present a collaborative decision model to study the energy exchange among building clusters where the buildings share a combined cooling, heating and power system, thermal storage, and battery, and each building aims to minimize its energy consumption cost under electricity demand uncertainty. The problem is formulated as a two-stage stochastic programming model and then solved using a hybrid decomposition algorithm that combines Sample Average Approximation algorithm with an enhanced Benders decomposition algorithm. Numerical experiments reveal that the hybrid decomposition algorithm provides high quality feasible solution in solving the realistic large-scale building cluster problem in a reasonable amount of time. Experimental results allow the investors to decide the optimal sizing of thermal and battery storage and PGU capacities under power demand uncertainty. Further, the model can assist decision makers to choose the appropriate pricing mechanism (i.e., an optimal pricing plan) under different electricity demand variability level. Finally, we observe that the CCHP-microgrid system is more sensitive to an increase in heating demand than the cooling demand.