A similar perverse incentive exists in the VC funding world that was parodied in HBO's Silicon Valley series-- namely an incentive against booking revenue early in the game. When you are "pre-revenue" then the mind can go wild with speculation, but as soon as you book revenue those become hard numbers and if they're unimpressive it turns people off. It's incredibly irrational of course, but investors are not on average more rational than other people.
In the ICO world an analog of this is that ICO projects that do ship inevitably suffer from hype deflation since the very first version of anything is never up to the hype that was used to sell it. So there's a dis-incentive to ship. Shipping means contact with reality, so failure to ship leads projects to wander off into scope-creep lala-land and never deliver anything.
I also think the large sums raised could cause pathologies that lead to project failure like ballooning scope, excessive hubris, over-hiring, or factional conflicts among founders. Too much money can defocus a project and too many new hires can actually reduce productivity.
Some ICOs are clearly frauds, but the rest I think are victims of these and other perverse incentives in the system.
Starting companies is hard. Rational investing is hard. Rationally investing in new companies is hard^2.