Mixed flow turbines have reached a level of maturity where iterative performance improvements are very small, with real performance benefits coming from better matching to a given application as opposed to improvements in technology. One ubiquitous design feature of mixed flow turbines used to control stress within the wheel is the radial fiber constraint, wherein blade material is stacked radially outward along the entirety of the blade. While this constraint yields a mechanical benefit, it constrains the aerodynamic design significantly, with the blade shape defined by one camberline. One potential means of realizing a performance improvement is the use of 3D blading, where the blade is not constrained to a radially fibered structure. In such a design, the blade shape could be freely modified to better control blade loading and secondary flows. This study investigated the viability of such 3D blading through optimization of a state of the art mixed flow turbine. An equivalent design was ensured by maintaining the meridional shape and operating conditions of the baseline (BL) wheel, thus facilitating a fair comparison between the radial and 3D wheels. The paper details the optimization including an innovative constraint-driven geometry modification tool, experimental validation of performance predictions, and an investigation into why 3D blading facilitated a performance improvement. The optimization process identified a performance improvement across the entire turbocharger operating line. With performance improvements facilitated through a reduction in tip leakage loss and improved pressure recovery within the conical diffuser. Importantly, the optimized design met targets for mass flow, maximum stress levels, and modal behavior, through the use of the novel geometry modification process.