The eradication of extreme poverty in fragile states is one of the central problems of global governance at the present time. Development of these states is hindered by instability, weak public and social institutions or ongoing conflicts and violence. The World Bank is a key partner of fragile states, which account for almost a third of the world’s population. This article is a continuation of research exploring the evolution of conceptual and practical approaches by the World Bank to cooperation with fragile states. Its methodology is based on a multilevel analysis of the securitization of foreign aid as proposed by J. Lind and J. Howell of the London School of Economics. The main focus of this examination is on the dynamics of the change of scale and structure of the World Bank’s aid to fragile states in comparison with global armed trends of providing aid to fragile states as well. This article concludes that statements about the priority of the Bank’s work in fragile states have not yet been realized in practice. The Bank remains committed to the standard approach to working with this group of recipients, which involves serious risks. The World Bank leans toward supporting projects in fragile states which increases volatility and reduces aid predictability. This trend undermines the development potentials of recipient states. Attention is drawn to political factors influencing aid flows to fragile states and particularly to the tendency of increasing the share of aid provided to fragile states through multi donor trust funds rather than through the mechanisms of the International Development Association (IDA). This trend indicates that the Bank is no longer a central point of aid distribution to the recipients, pointing to the lack of trust of donor states in the existing mechanisms and rules of aid distribution. It also reveals the expanding role of donors’ strategic interests in the process of choosing recipients of World Bank aid.