Some of that does happen now, but it always starts out as research. Many universities have a "industry partnership" program, either directly selling the research or starting companies to commercialize the research directly.
Research is the quest for new knowledge.
Engineering is the application of existing knowledge in creative and unique ways.
Being a good professor/researcher doesn't necessarily mean you are a good engineer.
Just like an architect may be good at conceptualizing a building --- but they will typically hire an engineer (often legally required) if they need to make sure the building can withstand a hurricane or an earthquake.
Particle accelerators are big tools, and the engineering is funded by grants. Ditto for telescopes. PyTorch allows thousands of labs to more efficiently do AI research, so should it have been fundable in some way?