Drones have been attracting the attention of diverse industries thanks to their superior maneuverability. Logistics companies especially keep trying to utilize drones for fast delivery following the growing market size of e-commerce. Accordingly, methods for safely operating multi-drone have been researched, and many researchers have proposed various optimal or near-optimal routing methods. However, such methods have some problems that cause routing failures or huge routing computation time in a drone-dense space due to many collisions. In this paper, we propose a centralized framework that deals with enormous collisions and obtains collision-free paths rapidly. We first build a drone energy consumption model with a data-driven method using flight experiment data of a commercial drone to estimate the drone battery state-of-charge (SoC). Then, we develop a novel routing method that generates collision-free paths by considering both the congestion of the space and the SoC of each drone. The proposed method is inspired by the VLSI circuit routing method that connects all signal nets among thousands of logic components. Through numerous delivery routing simulations, we confirm that the proposed method achieves a maximum of 6 times higher routing success rate with a 10x faster runtime compared with the state-of-the-art optimal method. In addition, we validate that the proposed method is applicable to delivery routing problems with various drone battery capacities.