Rentier states vary significantly in their commitment to and implementation of diversification programs, despite growing evidence supporting their economic necessity to sustain growth in fuel-rich countries. As early as the first oil boom in the 1970's, Saudi Arabia and Oman started designing and implementing diversification programs. While becoming increasingly dependent on natural resource revenues, other countries like Algeria and Libya have lagged behind in mapping out a diversification strategy. Given that diversification creates a trade-off between economic stability and potential political competition created by a shifting economic structure, what determines a rentier state's willingness to promote economic diversification and implement a long-term development strategy? If oil gives rentier regimes the latitude to pursue spending strategies to remain in power and secure their continued access to rent, what then determines the choice of such strategies? Because they are politically costly and difficult to implement, diversification programs can shed light on the political factors that determine allocation decisions in rentier states, and explain to some extent why some countries fail to avoid the resource curse. Yet the topic of diversification remains largely understudied, and little is known about the factors that lead some countries to implement such programs while others do not. This paper focuses on the intra-regime power dynamics that determine the observed variation in diversification patterns, and proposes a theory that links the structure of power in autocratic regimes to the adoption of diversification programs, conditional on the size of natural resources. Specifically, I argue that autocratic political power is structured at two levels determining the extent to which rulers will be inclined to promote diversification programs: (1) the state and non-state business elite level, and (2) the ruler's inner circle level. The configuration of power distribution at those two levels structures the incumbent's incentives to promote diversification. To test the theory, I compare diversification programs in Saudi Arabia and Libya between 1960 and 2010. With comparable oil qualities, large reserves, autocratic regimes, tribe-based societies and small populations at the time of oil discovery, Saudi Arabia and Libya have followed dramatically opposite diversification patterns that can be explained through the analysis of power distribution within each regime. By looking at the political factors that drive diversification strategies, this paper aims to uncover a powerful mechanism by which oil can affect the economic development and political stability in autocratic rentier states.