The occurrence of hypoxia in solid tumours is increasingly recognized as a limiting factor in the success of both radiotherapy and chemotherapy treatment, but at the same time offers a tumour-specific phenomenon for the activation of prodrugs. However, the design of clinically useful prodrugs that can be selectively activated in hypoxic cells has proved elusive. Specific reasons (activation by oxygen-insensitive two-electron reductases) have been proposed for the failure of quinone-based prodrugs, but a more general contributing factor may be inappropriate clinical trial design, and the failure to understand the critical importance of drug properties, such as efficient extra-vascular diffusion of the prodrug and back-diffusion of the activated drug in the tumour. Activation of prodrugs by therapeutic radiation and the use of hypoxia-selective gene therapy vectors, such as Clostridia, are exciting new mechanisms for prodrug research to explore, but are in much earlier stages of evaluation.