I don't think that no-holds-barred capitalism is the likeliest outcome for a superintelligent AI, but it is the one we need to be the most vigilant of. The abstract mathematician AI could just be allowed to live out its peaceful life proving ever-more-obscure theorems.