Integration of photovoltaic (PV) systems in modern smart grids has turned old energy networks into intelligent and integrated structures. Yet, this digitalization creates also new cybersecurity threats, since PV inverters, controllers and IoT-based sensors represent attack surfaces for attackers. In this work, we present an AI-based intrusion detection system for the detection and classification of cyberattacks against PV networks. The adopted architecture is a hybrid deep learning model, integrating CNN and RNN, to exploit features related to both spatial dependencies among network traffic and temporal dynamics for power-flow data. On benchmark intrusion datasets (CIC-IDS2017, UNSW-NB15) and injecting PV-SCADA legitimate traffic from simulation engine, the model has achieved detection accuracy of 97.2% and AUC score of 0.98 which significantly outperforms traditional machine learning algorithms such as SVM and Random Forest. It is also showing strong ability in detecting false data injection, denial-of-service and insider attacks with a low rate of false positives. In addition, deployment simulation in a smart grid environment demonstrates that the proposed framework is capable of real-time adaptive threat monitoring over distributed PV end host. Results validate the artificial intelligent as an effective method to improve cyber resilienc and operational reliability for smart solar facilities. The paper ends with a suggestion to introduce AI powered intrusion detection mechanisms through EMS and cybersecurity legislations in RE networks.