Commercial buildings spend vast amount of money every year on their electricity consumption. Though optimal scheduling and control of commercial buildings had been widely discussed in the literature, coordination of multiple buildings connected to power networks still need further consideration. Privacy issue is a major consideration to building managers, which makes the coordination problem very challenging. This article first provides a convex formulation for the optimal operation of heating, ventilation, and air conditioning systems in commercial buildings. Then, this article proposes the use of the alternating direction method of multiplier to enable privacy-based coordination of multiple commercial buildings in the electric grid. The formulated problem considers the thermal discomfort of the occupants and the building power consumption on the one hand, and the power losses of the power network on the other hand. The algorithm uses minimal amount of exchanged data to avoid any privacy invasion. A case study of multiple commercial buildings is done on an IEEE 15-bus distribution system to verify the proposed problem formulation. The results showed that the algorithm can achieve comparable performance to the centralized case in terms of minimizing the thermal discomfort, power consumption, and power losses.