Empirical evidence shows that secrecy in science has increased over the past decades, partly as a result of the commercialization of science. There is a good prima facie case against secrecy in science. It is part of the traditional ethos of science that it is a collective and open truth-seeking endeavor. In this paper, I will investigate whether secrecy in science can ever be epistemically justified. To answer this question, I first distinguish between different sorts of secrecy. Next, I propose an account of what it is for a practice to be epistemically justified, with the help of work by Alvin Goldman and Philip Kitcher. I then discuss motivations for secrecy in science that are found in the literature to see whether they amount to, or can be turned into, epistemic justifications for secrecy. The conclusion is that, although some forms of secrecy – particularly those motivated by universal moral concerns – are epistemically justified, secrecy that arises from special, often commercial, interests is not.