I'm not sure if it's necessarily big projects but the proving difficult theorems today has often involved the construction of huge "machinery", whole branch of math, that then get applied to simple-to-state-but-difficult-to-prove theorems, the example being Wiles using modular form theories to prove Fermet's last theorem.

And this situation comes because all (or the great majority) of the easy theorems have been proved for most established branches of math.

This also means great discoveries are coming at a later age for mathematicians, as simply getting up to speed in complex fields takes years.

All of this implies it would be hard to have another Von Neumann today.