The assessment of inhibitor risk is a crucial component in the clinical development of new and modified factor VIII (FVIII) preparations. There has been a recent discussion about the design of studies and the assessment of inhibitors and inhibitor risk in such studies at a recent FDA-sponsored FVIII Inhibitor Workshop, and new requirements for the success of these trials have been proposed to evaluate inhibitor data based on the use of an upper 95% confidence bound. We review the consequences of these requirements and demonstrate that for any product to succeed, it must have an extremely low underlying risk of inhibitor development. Furthermore, several existing commercially available FVIII products with an excellent safety record would not necessarily pass these endpoints. As a result, we propose an alternative set of acceptance criteria based on a Bayesian statistical paradigm. This approach is based on the determination of a probability that the product in question actually has an inhibitor risk below some pre-set limit, a concept that we believe is more intuitive than the traditional confidence interval method. We show that all existing products would pass this approach, but a product (Bisinact) with known inhibitor risk would not.