Uncertainties associated with cloud feedbacks are the largest among those contributing to equilibrium climate sensitivity (ECS) estimates by global climate models (GCMs). Here, we show that thermodynamic phase partitioning in mixed- phase clouds plays a fundamentally important role in cloud feedbacks in a doubled CO2 climate, and hence for ECS estimates, despite often being overlooked as an important aspect of cloud modeling. Liquid droplets and ice crystals not only influence the Earth’s radiative budget via their contrasting optical properties, but also through the effects of their lifetimes in the atmosphere, and hence play a critical role in shaping Earth’s climate. Using CESM1.0.5/CAM5.1 as the GCM of choice, the following six pairs of simulations (with present-day and doubled CO2 concentrations) were conducted: a control, one with an excessively high concentration of ice nuclei present in the atmosphere (“HighIN”), one with virtually no ice nuclei present in the atmosphere (“LowIN”), and two more with six different cloud microphysical tuning parameters in CAM5.1 adjusted in each simulation so as to most accurately reproduce observations of cloud phase obtained by NASA’s CALIOP instrument over a 79-month-long period (November 2007 to June 2014) in 20- degree latitudinal bands at the -10C, -20C and 30C isotherms. Calculations of ECS reveal that that of the HighIN and control simulations were lowest among all the simulations, with LowIN on the other extreme, nearly doubling that of the control simulation.