Does anyone know the collective time and energy put into deep learning models versus the social benefit? I recognize that this is near impossible to calculate

Does anyone know the collective time and energy put into deep learning models versus the social benefit? I recognize that this is near impossible to calculate and the benefits will hopefully be for many years to come. It does feel like the hype around deep learning has been large though and significant progress has been not as sticky as hoped.