Cloud computing (CC) is increasingly essential globally, revolutionizing collaborative services and data storage. However, it introduces heightened security concerns, notably with intrusion detection systems (IDSs) failing against sophisticated attacks. This study enhances IDS performance and efficiency in cloud environments using the CSE-CIC-IDS-2018 dataset. It involves data cleaning, exploratory analysis, normalization techniques (Z-score, min–max), feature selection via the Mother Optimization Algorithm (MOA), and a hybrid model, Batch Normalized Auto-Encoder with Gated Recurrent Unit (BN-AEGRU). Evaluating on updated cybersecurity datasets, it achieves up to 96% accuracy in binary and 98% in multi-class classification. The aim is to enhance intrusion detection efficacy and accuracy in cloud environments, contributing novel findings compared to existing literature.