Also the idea that US tech companies are doing much innovation is debatable. They are designing products for the international markets, but the core technologies are more and more coming from Asia.
Hence, great for software (and certain Schwerpunkte in hardware scaling) --- not innovation in general, but in fact solidifying a technological caste system: designers > SWE > HWE > every other kind of technical expert
Otherwise, you do need a Bell Labs to make sure your various experts are talking to one another and not constantly in fear of backstabbing by the management, their interns, other experts. (Especially considering in universities these days, rival groups in the same department have to make the same kind of calculations before cooperating)
I can see that YC does pay lip service to building community and such, but perhaps VC is too successful for its own good.
The key insight of venture capital was that firms like Bell Labs were holding on to very valuable resources at compensation rates that were far below what those resources could generate if they were empowered to create their own firms.
This was tremendously successful. While we have no counterfactual, innovation in the US blosomed ove the last 50 years. The fundamental research may have languished (I would probably disagree, moore's law didn't happen through magic for example) a tremendous number of companies, and all of the large companies we know of today, which provide all of the services of the modern world were a result of venture capital.