Second, Intel advances very specific, safe-bet science like shrinking the semiconductor. As plenty of other people have noted, the semiconductor was a public-funded invention and farmed off to Japan before it could be made "profitable" by American corporations. Heck, early American business laughed transistors off the continent before the value proposition was discovered.
Also, as long as I'm commenting here, I'll note that science is not dead, but I believe that our ability to consume information has tainted our ability to understand true science. To me, the quintessential American scientist of the 20th century was Richard Feynman. His Cargo Cult Science [1] essay, which pops up on HN from time to time explains what's happening best: people who claim to follow science, are actually practicing some sort of modern-day witch doctor magic. They wave their hands and claim the data supports their wild discovery and someone bites. Suddenly it takes 10,000 hours to become a master of something "on average."