Utility functions and such? Or, even without trying to cram behavioral economics into a more traditional framework. At some level, it may be rational to trade off some measure of certain income into a big (but very hard to define percentages of) bet in terms on money/being in at ground floor of something great.

It may be a sucker bet in a lot of cases. But expected value (given how hard that may be to compute for unicorns) isn't always the best bet either given vast uncertainties.