When compared to other nuclear weapon powers why has India historically lagged in the development, deployment, and operational planning of its nuclear force despite unambiguous national security threats? My dissertation answers this question through a cross-sectional study of three decades of Indian nuclear decision-making from 1980 until 2010. Using a multi-disciplinary approach, which combines insights from New Institutionalism, organization theory and Cognitive Psychology with historical process tracing and elite interviewing methods, I argue that there are two interrelated causes for the Indian state’s historic underperformance: (a) the absence of a strongly institutionalized “epistemic community” within the state; and (b) the absence of shared policy-planning and decision-making processes. The first cause is institutional while the second is organizational. I show that epistemic communities as knowledge brokers are necessary for socializing a state’s decision-makers into new learning practices. For learning to occur, epistemic communities must also operate in relatively open and non-monopolistic policy planning and decision-making environments. The latter reduce the scope for heuristics and cognitive biases and are conducive for relatively rational and optimal policy outcomes. I present evidence to show that Indian decision-makers partially mobilized a national security-centric “epistemic community” in the pre-1998 era; and only slowly institutionalized it within the state in the post-1998 decade. These base conditions when grafted on to highly centralized, compartmentalized and monopolistic policy planning and decision-making processes, attenuated the Indian state’s policy capacity. The net result has been policy outcomes riddled with heuristic and cognitive biases alongside the weak actualization of instituted policies.