Wireless sensor network (WSN) architectures involve multiple sensor nodes that cover large areas. It is difficult to deploy such nodes efficiently. However, based on clustering and collaboration, WSNs can organize efficiently and prolong system lifetime. In this paper, we discuss the design and implementation of a collaborative LoRa WAN using ultra-low-energy hardware and smart utilization of battery power. We have chosen the robust LoRa modulation technique and clustering method to design the end nodes. The sensor data from each of the end nodes are transmitted using low-power, long distance, and noise-resilient LoRa radio channels to their corresponding Cluster Heads(CHs), where it is first stored and then processed before transmitting further to the Gateways. The storing and then transmitting technique reduces the bandwidth requirement significantly. Also, the stored sensor data can be analyzed for system optimization. Immediately after transmission, the microcontroller goes to hibernation mode, whereas, the sensor system and the LoRa module are powered off until the next cycle. In this way, the power consumption by the end node is kept at its minimum and the battery life is extended to 35 years. Several challenges related to hardware design and network parameters optimization to achieve energy-efficient IoT-based LoRa WAN are presented and discussed.