Recent work developed a novel, biomimetic, cost-effective three-dimensional hollow fibre bioreactor for growing healthy red blood cells ex vivo (Panoskaltsis et al., 2012). This bioreactor recapitulates architectural and functional properties of erythrocyte formation and thereby reduces the need for expensive growth factors by more than an order of magnitude. Individual experiments to empirically improve the bioreactor are intensive, so we propose global superstructure optimisation for bioreactor design. Our approach integrates topological design choices with operating conditions. Design choices include: number of parallelised bioreactors; number and type of hollow fibres; size and aspect ratio. Operating conditions are: feed concentrations; flowrate through the reactor. This manuscript quantitatively demonstrates, for the first time, the potential for ex vivo red blood cell production to compete openly against the transfusion market for rare blood. We discuss the potential of superstructure design not only on this individual bioreactor but also more generally on bioprocess optimisation.