What could government invest in to earn an exponential return?

What could government invest in to earn an exponential return? Early childhood education? Preventive medicine? Other ideas?

Assuming agreement can be achieved politically, we need to note that it is hard to quantify a real return on investment. But to mildly qualify our assumption, while it is true and demonstrably so (thus even quantifiable) that every dollar invested in early childhood education or preventive medicine offers an exponential return, the return is in money saved, not earned by governments.

Governments do not ever save money, they just _may_ collect "less" tax revenue, thus giving tax payers the chance to pay "less" taxes for (presumably) the same or better government services. The last part gets so fuzzy, that it is rarely if ever possible to assign weights to factors having determined the changes in taxation policies or budget priorities alone, so forget about tracing back "savings" as ROIs.

Shifting money spent on prisons and policing and welfare to spending it on quality of life improvement projects becomes the return on investment in early childhood education that we all may seem to want to suggest (or so I hope--based on a Lincoln quote I love about the power of our children in shaping their future).
The same with preventive medicine. Healthy people make a happy and productive workforce. People advance in careers, have increased incomes which get taxed more, while also making employers better profits/profit margins which again get taxed more, but which also allow for further investments and thus sustained economic growth, and the government again taxes more...

It's been said that Internet (back in the dial-up era) was one of these fields we are now asking about. Now the hype has been for a while broadband in rural areas. I may agree and disagree with both. I disagree because wires or fiber optic do not change anything. Instead, _services_ available over the medium do. So if broadband in rural areas improves the equity of access to education/knowledge or (preventive) health care of rural residents, then the technology would qualify under a positive for your question. Yet, if we stop at deployment of fiber alone and do not match said investment with an even larger one in making new targeted services available to beneficiaries, that investment in broadband alone becomes almost meaningless in terms of exponential government ROI.

The problem, as always and as everyone most likely knows better than I do, is that often government officials are usually naturally reluctant to tackle long term policies or problems, as on those it is always hard to see immediate results while in office or seeking reelection. Federally sponsored R&D in pharmaceuticals has been among those you ask about for years now, yet depending again whom you ask. Many military technologies later declassified and adapted to civilian use are among those too. We all read and write over the grand-grandchild of a former DARPA network turned the Internet, with huge exponential government ROIs, which have even enabled President Clinton to balance the budget during the .com bubble of the late '90s. How more quantifiable than that can we conceive something to be? Before that satellite tracking technology, again civilian disseminated in the meantime (as almost everyone has a GPS in their pocket/car these days--those firms get taxed), and did I say airplanes? For today bioengineering, human genome research etc. the list is too long...

It may be easier to quantify ROIs on the latter cases, past, present, or future.