The NIH is finding this out with their recent push to turn NIH-funded results into drugs. Collins is already backpedaling from earlier this year (you'd think he would have learned this lesson after being beaten to the genome by an organization of 1/10th the means).
At Apple, I worked closely with battery engineers. Lots and lots of battery research is going on -- we hear about nano-virus-assembled anodes all the time. But approximately none of it is useful to battery engineers. Batteries must have energy density, power density, good recharge time, and high cycle life. Press releases about gee-whiz battery chemistries typically quote one or two of these figures, but never all four. That's because all of them have a huge problem with one or more of these things, and no horizon for fixing them. And the instant a breakthrough is made -- like graphite anodes at Bell labs in the early '80s -- everyone will be working on it. There are approximately no 'forgotten breakthroughs' in human society. The incentives provided by contributing to wikipedia are enough to disclose almost everything known.
The whole point of funding basic research is to discover such things, which are too important to be proprietary. Markets are very good at the details that are appropriate to keep under one roof.
It's rarely a straightforward thing to do, and I can understand academics not wanting to tackle it. But it really does matter.