(abridged) The Atacama Large Aperture Submillimeter Telescope (AtLAST) project aims to build a 50-m-class submm telescope with $>1^circ$ field of view, high in the Atacama Desert, providing fast and detailed mapping of the mm/submm sky. It will thus serve as a strong complement to existing facilities such as ALMA. ALMAs small field of view ($<15^{primeprime}$ at 350 GHz) limits its mapping speed for large surveys. Instead, a single dish with a large field of view such as the AtLAST concept can host large multi-element instruments that can more efficiently map large portions of the sky. Small aperture survey instruments (typically much smaller than $<3times$ the size of an interferometric array element) can mitigate this somewhat but lack the resolution for accurate recovery of source location and have small collecting areas. Furthermore, small aperture survey instruments do not provide sufficient overlap in the spatial scales they sample to provide a complete reconstruction of extended sources (i.e. the zero-spacing information is incomplete in $u,v$-space.) The heterodyne instrumentation for the AtLAST telescope that we consider here will take advantage of extensive developments in the past decade improving the performance and pixel count of heterodyne focal plane arrays. Such instrumentation, with higher pixel counts, has alredy begun to take advantage of integration in the focal planes to increase packaging efficiency over simply stacking modular mixer blocks in the focal plane. We extrapolate from the current state-of-the-art to present concept first-generation heterodyne designs for AtLAST.