I seem to remember that von Neumann was also responsible for what later came to be known as "Mutually Assured Destruction", the idea through game theory that ensuring that both nations would be prevented from acting preemptively because the results would carry too great a cost.
Their predictions on nuclear arsenals were correct. A nuclear war in 1965 would have been vastly worse than one in 1950.
What they got wrong is that great powers would successfully avoid war over the long term. Which at the time was a bet with very poor odds.