Energy generation, sources and distribution methods have been continuously evolving over the past decade. With the increased efficiency associated with solar energy production and distribution, local homeowners have also assumed the role of energy generators, even getting credit for access electricity supplied to the grid given the policy around net-metering. When planning their energy distribution frameworks, electricity providers have to take these changes in energy consumption and generation into account. However, little is known about how solar energy systems impact the demand and supply of grid electricity managed by utility companies. This study proposes a new approach to solar energy predictive modeling which combines machine learning and a variety of publicly available data sources to predict site-specific temperature and solar irradiance (the two primary “missing ingredients”). The preliminary findings show a decreased error when using the new approach (near-future data) in comparison to the traditional approach (historical data) for predicting solar energy generation. As the adoption of solar energy increases, so will potential disruptions to the grid. These preliminary findings show the potential for aggregating individual site-specific predictions to the regional level for the purpose of estimating area-specific solar energy disturbances and moving efforts towards predictive grid optimization.