You just described in a nutshell why it's usually governments that make big infrastructure-scale and basic research investments.
Governments either don't care about ROI at all and are motivated by other factors (war, national security, national development, political points, etc.) or can afford to look at ROI over exceedingly long time scales.
However, governments are piss-poor investors. So what we get is private capital -- which invests in a lot of tiny incremental things -- and a lumbering beast called "public capital" that invests in big things in a clumsy, ineffective way.
Governments either don't care about ROI at all and are motivated by other factors (war, national security, national development, political points, etc.) or can afford to look at ROI over exceedingly long time scales.
However, governments are piss-poor investors. So what we get is private capital -- which invests in a lot of tiny incremental things -- and a lumbering beast called "public capital" that invests in big things in a clumsy, ineffective way.