Falls can cause severe injuries to elder people. Many studies have been conducted in order to develop devices that improve the recognition of fall events, while very few have demonstrated approaches for the minimization of their power consumption. We propose a hardware/software co-design of an Internet of Things (IoT) fall detection device that takes advantage of accelerometer’s embedded functionalities and enables its interrupt-driven operation while the rest of the circuit is in shutdown mode. In contrast with most low power fall detection devices, the proposed device supports Wi-Fi connectivity that enables ubiquitous use and real-time remote fall events monitoring through cloud services, without the need of external equipment or a local server. The device integrates a vibration motor and a cancelation button, which notify the detection of a fall event and waits for Patient’s response, while the implementation of battery monitoring mechanism enables continuous device use.