Metal cutting fluids for improved cooling and lubrication are an environmental risk and a health risk for workers. Minimizing water consumption in industry is also a goal for a more sustainable production. Therefore, metal cutting emulsions that contain hazardous additives and consume considerable amounts of water are being replaced with more sustainable metal cutting fluids and delivery systems, like vegetable oils that are delivered in small aerosol droplets, i.e. via minimum quantity lubrication (MQL). Since the volume of the cutting fluid in MQL is small, the cooling capacity of MQL is not optimal. In order to improve the cooling capacity of the MQL, the spray can be subcooled using liquid nitrogen. This paper investigates subcooled MQL with machining simulations and experiments. The simulations provide complementary information to the experiments, which would be otherwise difficult to obtain, e.g. thermal behavior in the tool-chip contact and residual strains on the workpiece surface. The cBN hard turning simulations and experiments are done for powder-based Cr-Mo-V tools steel, Uddeholm Vanadis 8 using MQL subcooled to -10 °C and regular MQL at room temperature. The cutting forces and tool wear are measured from the experiments, that are used as the calibration factor for the simulations. After calibration, the simulations are used to evaluate the thermal effects of the subcooled MQL, and the surface residual strains on the workpiece. The simulations are in good agreement with the experiments in terms of chip morphology and cutting forces. The cutting experiments and simulations show that there is only a small difference between the subcooled MQL and regular MQL regarding the wear behavior, cutting forces or process temperatures. The simulations predict substantial residual plastic strain on the workpiece surface after machining. The surface deformations are shown to have significant effect on the simulated cutting forces after the initial tool pass, an outcome that has major implications for inverse material modelling.