The Dark Energy Survey has provided leading tests of our cosmological model with gravitational lensing-based measurements of structure. I will review these and highlight the fact that uncertainty in galaxy redshift distributions are limiting present, but especially future, such studies. For the DES three-year analysis, we are developing a novel method that uses the high-quality and multi-color (ugrizYJHK) information of deep photometric fields as an intermediary between wide survey photometry and spectroscopic samples, greatly reducing cosmic variance in redshift distribution calibration. I will discuss the prospects of this, and the requirements for additional deep-field and spectroscopic data sets for the success of DES and LSST.