I think you're missing the point of the article. $300M is also in the ballpark for a satellite. Presumably communications companies wouldn't spend that money to launch a satellite unless they had a pretty good indication of a massive revenue stream at hand. That doesn't make it wrong. (Doesn't make it right, either.) And really, that example cuts the opposite direction from what you're suggesting. Spending $300M to build out your infranstructure is never "risk-free". What if market conditions are such that trading falls and they can't generate as much revenue as they thought? What is somebody spent $301M and built a better cable mousetrap?