<p>In deep-ocean human submersibles, effective management of on-board oxygen, energy and human cabin microclimate are essential for survivability of crew during emergency situations. This article describes the development and validation of Prana and Garuda modules of Situation-aware Cognitive Digital Twin Chaitanya, based on machine-learnt ocean environment, battery performance under varied temperature conditions, Matsya’s hydrodynamics, crew physiology and thermal model of human cabin. As a digital co-pilot, Prana and Garuda modules of Chaitanya supports the mission with advisories for maintaining conducive human cabin microclimate, ensuring availability of propulsion power during delayed retrievals and optimizing oxygen usage for crew survivability beyond emergency periods. The novelty of this work lies in the development of a situation-aware Cognitive Digital Twin that integrates machine-learnt ocean environment prediction, thermal–physiological cabin modelling, hydrodynamic behavior, and battery performance under temperature variations. Unlike existing literature that treats life-support, hydrodynamics, and energy systems independently, the proposed Prana and Garuda modules operate as AI-enabled digital co-pilots that jointly optimize oxygen usage, cabin microclimate, and propulsion survivability during emergency retrieval delays. The oxygen consumption model of Prana is found to match the crew oxygen consumption during surface floating condition logged during the field tests with an accuracy of 96% and is used for training the Cognitive Digital Twin. Simulation results of Garuda indicate that residual propulsion energy of 20 kWh could support 64 h of subsurface hovering at 220 m water depth in Central Indian Ocean.</p>