Implantable cardiac devices are crucial to the management of cardiac disorders. Nevertheless, their short battery life remains a critical challenge owing to the need for frequent replacement, which increases the risk of surgery. We investigated the potential of lightweight artificial intelligence (AI) models to optimize the energy consumption of these devices. Using publicly available datasets (PhysioNet, MIT-BIH Arrhythmia Database) in a simulation-based evaluation, our approach achieved up to 86% energy savings while maintaining arrhythmia detection accuracy above 98%. These findings suggest that lightweight AI models can significantly extend battery life without compromising diagnostic performance, paving the way for more sustainable and patient-friendly implantable cardiac devices. Further validation on physical hardware and in clinical settings is warranted. This project opens up perspectives on how AI could help improve the efficiency and extend the longevity of implantable cardiac devices while delivering more sustainable healthcare solutions.