Future smart grids will require a flexible, observable, and controllable network architecture for reliable and efficient energy delivery under uncertain conditions. They will also necessitate variability in distributed energy generators and demand-side loads. This study presents a tree-like user-mode network architecture responding to these requirements. The approaches presented for the next-generation grid architecture facilitate the management of distributed generation strategies based on renewable sources, distributed storage, and demand-side load management. The authors draw a framework for the future digital power grid concept and assess its viability in relation to volatile, diverse generation and consumption possibilities. In this sense, probabilistic energy balance analyses of tree-like user-mode networks with a stochastic end-user population are conducted to investigate the energy reliability of the proposed grid. A case study based on published generation and consumption profile data is also presented, and several generation scenarios formed by variants of this data are discussed. (c) 2012 Elsevier Ltd. All rights reserved.