This paper investigates the potential of using aggregated heating, ventilation, and air-conditioning (HVAC) loads to smooth out the power fluctuation of a wind farm and/or a solar farm to make these stochastic resources more dispatchable. Specially, we consider a Virtual Power Plant (VPP) which consists of several wind/solar power units, a number of buildings with their HVAC systems, and a battery storage facility. A portion of the power generation from renewable energy sources (RESs) is used to operate HVAC systems and the rest (if any) is sold to the main grid. The design goal is to determine an optimal hourly scheduled power dispatch that the VPP must submit to electricity market to maximize its benefit. The short-term fluctuation of renewable energy generation (i.e., over intrahour intervals) is mitigated by smartly adjusting the flexible HVAC load, which enables the VPP to provide a firmed hourly dispatch. The underlying optimization problem is formulated as a two-stage stochastic program where system uncertain parameters are modeled using Monte-Carlo simulation. Building thermal dynamics model and users' climate comfort criteria are considered in the proposed optimization framework. Numerical results is presented to illustrate the effectiveness of the proposed model.