Roller element bearings present in the intermediate and high-speed stages of wind turbine gearboxes operate in dynamic working conditions and in some cases may fail within 30% or less of their designed lifetime. Upon investigation, it has been identified that these premature failures happen due to a peculiar failure mode associated with formation of white etching cracks (WEC). This continues to be a great challenge for the wind energy operators as it leads to an increase of maintenance and operation costs in addition to long wind turbine downtime. Therefore, the industry is in dire need of a lifetime prediction methodology that could take in multi-scale inputs ranging from bearing loads at the system level down to the level of bearing material properties at the microscopic level. This work summarizes the overall approach of a project that aims towards an integrated framework which links load data from the bearings and microstructure related non-metallic inclusion statistics from bearing steels, to predict a material based probability of failure. The interlink between both aspects is a numerical rolling contact fatigue (RCF) framework based on finite element analysis, which includes multi-scale data as an input to calculate rolling contact fatigue damage. The outcome will help the wind industry to better predict bearing failures.