Eliminating Bayh-Dole would mean university-based patents generated with taxpayer funds would be available to any interested party under a non-exclusive licensing program, and then corporations would again be incentivized to maintain private research centers - which IIRC also served a tax-writeoff function for AT&T in Bell Lab's heyday.
At least some of the 80s vintage corporate labs like DEC were being increasingly folded into the broader engineering organizations as their parents became less dominant. Essentially corporate research labs are a creature of organizations that have some long-term play money. Or at least that's the idea. I'd note that some of the current work going on in quantum computing has a lineage that dates back to some pretty fundamental research done in corporate labs in the 1960s.
Currently leading tech research? Google, Meta, Microsoft, Amazon, NVIDIA, Intel, and for some ungodly reason, Elon Musk with his government grants and Teslabucks. These are all effective monopolies in several sectors to the point where dethroning any of them is going to require either a massive change in the world through new technology or extensive multi-governmental legal action. These companies have the ability to spend a lot of money on something stupid (for the ones we have seen in the public eye, see Stadia, Metaverse, Itanium, we had Plan 9 as Bell's last hurrah into the void, these types of stupid things.) which loses it for the promise of an eventual golden egg. As a theory I'd say this is a pretty good one considering that it extends well into the modern day.
Ideally, we'd probably fund research in some better way, but I have no clue how the hell that would even work. Until then, the future comes in the form of price fixing and anti-competitive practices.
In general, the US probably does a decent job of funding research. Small scale research orgs within companies are often practical. There are company and government grants to both academia and commercial/academic partnerships. VCs, for their faults, do fund pretty speculative work.
For the reasons you say, big corporate research labs require conditions that are mostly not an unalloyed good at best.
> "For over half a century, Bell Labs had been owned by the telephone monopoly AT&T and had plenty of money to spend on science. But in 1984, the monopoly broke up, and after 1989, managers encouraged Bell Lab researchers to focus on research with commercial applications. Disenchanted, top scientists began to leave for universities and were mostly not replaced by new recruits. In 1995, ownership of Bell Labs was transferred to the newly formed company Lucent Technologies."
Then Lucent got hit by the dot-com bubble blowout, resulting in a 30% loss of share value in Jan 2000. This led to a new focus on PR:
> "By keeping up its practice of releasing exciting scientific findings, the lab could continue to demonstrate to investors, customers and anyone else that Lucent had a sound, long-term technological future. Again, this was a question of survival: with revenues falling, managers had to make the argument that their jobs and the jobs of their staff were worth keeping."
This led to a culture in which critical scrutiny of the claims of one fraudulent researcher was discouraged in favor of institutional cheerleading, and the end result was that 15 papers published in Science and Nature had to be retracted.
[1] "Plastic Fantastic: How the Biggest Fraud in Physics Shook the Scientific World" (2009) Eugenie Samuel Reich.
If you want a parable for what's happened to Bell Labs and many other scientific institutions in the US, it's that of the greedy farmer killing the goose that laid the golden eggs.