Well, it's my understanding that most prize money won by scientists are not reinvested in their research, meaning that it act more as a regular enterprise bonus than a kind of research grant. Is it false?

In that case, I think it was proven than passing a certain mark, the amount of money in a prize get diminishing returns quickly, thus the proposal that it's more interesting to give to a lots of scientists a small sum (not necessarily $3000, that was derived from the $3 million) than for a few a big one.