> Assuming your figures are correct, you'll get double the data for another 30% cost.

The debate over the cost to build another is largely moot. You'd struggle to find the people to do it for many reasons. The talent is even scarcer than the money.