The Euclid, Rubin and Roman projects will undertake flagship optical/near-infrared surveys in the 2020s. They will map thousands of square degrees of sky and cover the electromagnetic spectrum between 0.3 and 2.3 microns with sub-arcsec resolution. Joint survey processing (JSP) is the pixel level combination of the Rubin/LSST, Euclid, and Roman datasets. By combining the high spatial resolution of the space-based datasets with deep, seeing-limited, ground-based images in the optical bands, systematics like source confusion and astrometric mismatch can be addressed to derive the highest precision optical/infrared photometric catalogs. This enables a range of time-domain, cosmological and astrophysical science by the community ranging from solving for asteroid trajectories and composition, to epoch of reionization studies, and the nature of dark energy. This talk is about the science drivers for optimally combining the 100 Petabytes of reduced data from Euclid, LSST and Roman and building the computational and networking infrastructure to enable distributed co-processing of these voluminous datasets.