Abstract. It is argued that it is far more cost effective to carry out some projects with medium-sized dedicated zenith telescopes rather than large steerable telescopes, freeing the later to carry out projects that truly need them. I show that the large number of objects observed with a surveying 4-m zenith telescope allows one to carry out cosmological projects at low redshifts. Examining two case studies, I show first that a variability survey would obtain light curves for several thousands of type Ia supernovae per year up to z = 1 and easily discriminate among competing cosmological models. Finally, I discuss a second case study, consisting of a spectrophotometric survey carried out with interference filters, showing its power to discriminate among cosmological models and to study the large-scale distribution of galaxies in the Universe.