This paper deals with a DC-segmentation of interconnected power systems which represents a relatively new and interesting approach that could be a potential solution for mitigation of the cascading outages effects. The DC-segmentation is defined as the use of DC links to split large AC grids into sets of smaller asynchronously operated AC segments. By implementing grid segmentation, it should be plausible to improve the controllability of power flows, to limit the propagation of disturbances, to reduce the risk of cascading outages and blackouts and to decrease the complexity of power systems. The statistical analysis of time series of blackouts size measures demonstrates the power law region in distributions of different quantities. This heavy tail behavior implies that the blackouts of different scales may take place and the extreme events cannot be overlooked. In order to take into account the risk of large blackouts, a new risk assessment method is developed and implemented. This method is based on Monte Carlo Simulation and takes into account the cascading outages of transmission system due to the overloading and the hidden failure of protection systems and the response of generators/loads/HVDC controllers to the power imbalance or the frequency deviation. The effectiveness of the DC-segmentation for reducing the risk of large blackouts is assessed using the proposed risk assessment method applied to a simplified model of the potential future intercontinental power grid of Europe and Russia. The distributions and expected values of the blackout size quantities (including Load Not Served (LNS), Number of Transmission component Outage (NTO) and Energy Not Served (ENS)) are evaluated before and after the DC-segmentation.