<p>The accurate subseasonal prediction of the spatio-temporal rainfall variability of the Indian summer monsoon holds immense socio-economic importance. During the 2022 monsoon, slightly above-normal rainfall was recorded across most regions of India, while below-normal precipitation was observed in the eastern parts, particularly over the Indo-Gangetic plains. In this study, the real-time performance of a recently developed multi-physics multi-ensemble (MPME) subseasonal prediction system is evaluated for its ability to predict the spatio-temporal rainfall variability during the 2022 monsoon. Subdued rainfall activity in June and an asymmetric rainfall distribution, characterized by a robust positive pattern over northwestern regions and a strong negative pattern over eastern and northeastern India in July and August, were reasonably reproduced by the MPME system. Additionally, the role of extra-tropical intrusions was investigated, focusing on the formation, propagation, and dissipation of West and East Asian blocking ridges, which were associated with the pronounced east-west rainfall asymmetry over India. The potential of the MPME system to forecast active and break spells 2–3 weeks in advance is demonstrated. The intra-seasonal oscillation (ISO) was identified as the dominant driver of rainfall variability during the 2022 monsoon, with the model exhibiting higher monsoon ISO (MISO) prediction skill in 2022 compared to its hindcast skill (2003–2018). Overall, the MPME prediction system was found to be skilful over most of the homogeneous regions of India up to a three-week lead time. However, decay in skill was noted beyond two weeks over smaller spatial domains, such as meteorological sub-divisions.</p>