Artificial Intelligence (AI) is revolutionising the healthcare sector and can be used to assist clinicians in decision-making. It can be used to reduce the workload of radiologists and positively impact the clinical workflow. To generate radiology reports, this work introduces a novel architecture, MoERad: Mixture of Experts for Radiology. MoERad combines a context-aware pre-trained Convolution Neural Network (CNN) and a Mixture of Experts (MoE) model to interpret Chest X-ray images and generate radiology reports. To our knowledge, this is the first work that implements MoE for automatic radiology report generation tasks. The pre-trained CNN helped the MoERad learn fine-grained details from the data during training, performing better on unseen data. A comprehensive ablation study confirms the benefits of using a pre-trained CNN and highlights the importance of the MoE block over traditional dense MLPs. MoERad has performed significantly better than the current state-of-the-art models on the ReXrank leaderboard. MoERad has achieved top performance on 4 out of 8 evaluation metrics on the ReXGradient dataset. While evaluating against the IU-Xray dataset, MoERad has achieved top performance on 5 out of 8 evaluation metrics, confirming its robustness and generalisability. The code for this work can be found in https://github.com/gssriram/MoERad .