My greatly advanced age allows me to comment. I got my PhD in computer science in 1983, and spent some time in research, moving quickly to "industrial research", and so I was very aware of Xerox PARC and similar institutions.
- Xerox PARC was, of course, famous for cutting edge computer science research that could have, and often did impact the computing world, although sometimes years later, and never to the benefit of Xerox. I don't think this was by design, but it sure worked out that way.
- Bell Labs and its impact needs no introduction here. They are not known for impacting AT&Ts bottom line, although if it did, I would not be aware of it.
- IBM Labs did some great work, and probably had the most successful research to product pipeline. I'm thinking of PL/1 and DB/2 mainly, but I'm sure there are many other examples that I'm forgetting. APL might also be an example. I suspect that much of their mainframe OS work benefitted from this pipeline, but I can't comment since I had little exposure to mainframe OSes.
- Microsoft is a good example too. I think they have had some success feeding research into product. Or maybe it is more accurate to say that they had some success in hiring researchers to work on products. (I came very close to being one of those researchers, but decided to go into startups instead.)
- A far more obscure example: I worked at Computer Corporation of America in the 80s. At the time, they were one of the great database research labs, (e.g. Bernstein and Goodman on concurrency control), and I was lucky enough to work there for a while. CCAs main product was Model 204, an ancient, very fast, hierarchical database system that ran on mainframes, and was, I believe, coded in Assembler. CCA also had a consulting side, that did DARPA-funded research, and that's how they funded database research. Profiting just from that research did work, for a while. Then it stopped working, and a past-its-prime Xerox PARC purchased a past-its-prime CCA research group (if I'm remembering correctly). At that point, I left to join an object-oriented database startup (remember those?). And sure enough, some of the ideas I worked on at CCA (adding persistence to Ada) did show up in our product. So again, research to product, but not benefitting the company itself.
Now? I think that there is cutting-edge research at Google and Amazon, and both are excellent examples of fundamental research leading to immediately useful products.
