NSF has actually been experimenting with this sort of funding model for a couple years through its new (as of 2019) convergence accelerator program, which I think is awesome. It’s explicitly a multi-sector program where academics partner with companies to do some proof of concept research in a phase 1 award and translate that into a viable commercial product or innovation in a phase 2 award. Potential for long term return on investment is explicitly part of the review criteria, which target this sort of historically underfunded middle ground between basic research and technology development

https://www.nsf.gov/funding/initiatives/convergence-accelera...