The discovery of billion solar mass quasars at redshifts of 6-7 challenges our understanding of the early Universe; how did such massive objects form in the first billion years? Observational constraints and numerical simulations increasingly favour the “direct collapse” scenario. In this case, an atomically-cooled halo of primordial composition accretes rapidly onto a single protostellar core, ultimately collapsing through the Chandrasekhar-Feynman instability to produce a supermassive (~100,000 solar mass) “seed” black hole. In this talk, I’ll present a systematic study of the lives and deaths of these objects, using the 1D implicit hydrodynamics and stellar evolution code KEPLER. We include post-Newtonian corrections to gravity and a detailed treatment of nuclear burning processes using an adaptive network. We find a simple relation between the infall rate and the final mass at collapse, and rule out the existence of rapidly-rotating supermassive stars. I’ll also discuss the possibility of early chemical enrichment from these objects, observational prospects in the era of the JWST, and briefly summarize other future directions agreed upon at our workshop “Titans of the Early Universe” held at the Monash Prato Centre in Italy, in November of last year.