The late founder had supported the direction of the research. It's just that the new CEO does not have the same level of risk taking vision, and instead listened to prevailing experts on the direction. He had assumed that this lone researcher is just following a dead end.

Childish behaviour aside, the conservative decision to cut costs and shut down the research is justifiable to some degree. The only problem is that the GAN research path is well trodden by others already, and therefore, a smart CEO should see that it is also a dead end.