Very few supercomputers on Earth have that sort of power. What sort of problems could you apply that threat to? We deal in DoS attacks from thousands of PCs - but millions? Most risk models today assume a finite capability for the adversary, but this completely changes that.
