Development of large, far-infrared telescopes in space has taken on a new urgency with breakthroughs in detector technology and recognition of the fundamental importance of the far-infrared spectral region to cosmological questions as well as to understanding how our own Solar System came into being. SAFIR is 10m-class far-infrared observatory that would begin development later in this decade to meet these needs. Its operating temperature (less than or equal to 4 K) and instrument complement would be optimized to reach the natural sky confusion limit in the far-infrared with diffraction-limited peformance down to at least the atmospheric cutoff, gimel greater than or equal to 40mum. This would provide a point source sensitivity improvement of several, orders of magnitude over that of SIRTF. SAFIR's science goals are driven by the fact that youngest stages of almost all phenomena in the universe are shrouded in absorption by and emission from cool dust that emits strongly in the far-infrared, 20mum - 1mm. The earliest stages of star formation when gas and dust clouds are collapsing and the beginnings of a central star are taking shape can only be observed in the far-infrared. Likewise, the cool dust that will eventually form planetary systems, as well as the cool "debris" dust clouds that indicate the likelihood of planetary sized bodies around more developed stars can only be observed at wavelengths longward of 20mum. The most active galaxies in the universe appear to be those whose gaseous disks are interacting in violent collisions. The details of these galaxies, including the central black holes that probably exist in most of them, are obscured to shorter wavelength optical and ultraviolet observatories by the large amounts of dust in their interstellar media. Early stages of galaxy formation appear to result in powerful sub-mm emission indicative of substantial metal enrichment early in the history of the universe. Finally, the warm gas of newly collapsing, unenriched galaxies should reveal itself in hydrogen emission at these long wavelengths. The combination of strong dust emission and large redshift (1 < z < 5) of these galaxies means that they can only be studied in the far-infrared and sub-mm where SAFIR will provide the sensitivity and angular resolution to peform imaging and spectroscopic studies of individual galaxies in the early universe. The main drivers on the telescope are operating temperature and aperture. SAFIR can take advantage of much of the technology under development for NGST. Because of the much less stringent requirements on optical accuracy, however, SAFIR can be developed at substantially lower cost.