How do unbalanced nuclear alliances provide extended nuclear deterrence (END) to their members? Why have nuclear alliances chosen certain types of END strategy and not others? Existing accounts regard END as a function of the inter-alliance balance of power, regime type, or institutional design. END strategies inspired by theories focused on regime type and institutional design have not yet materialised, while the inter-alliance balance of power does not suffice to explain the choice of END strategy. To elucidate variations in END strategy, this article puts forward an argument centred on the intra-alliance balance of power. Drawing on the history of the US-led and the Soviet-led alliances during the Cold War, namely North Atlantic Treaty Organization and the Warsaw Pact, it shows how the two superpowers changed their approach to defending their allies with nuclear weapons according to quantitative and qualitative shifts in the distribution of power within the alliance.