Are you saying in a perfect world, the perfect CEO would "can" (get rid of) every team/project that isn't projected to 5% ROI yearly (assuming risk free rate has roughly gone from 0.25% -> 5%)?
Obviously there are exceptions where you want to eat losses up from (R&D a big project for a few years hoping it returns a lot down the road)?
How does this apply given that we're assuming Google has enough cash on hand from profit of other aspects of their business that they do not finance most projects through new loans (debt) from banks at current interest rates because they don't need to?
Their cash balance doesn't really matter, it impacts their cost of capital in various ways, but you can still compare the ROIC of any given project to the overall company's cost of capital.
Obviously there are exceptions where you want to eat losses up from (R&D a big project for a few years hoping it returns a lot down the road)?