You're begging the question; there is a country beyond Silicon Valley, and Grove's argument is that what is best for Silicon Valley (meaning, the bottom lines of a few tech firms which together only employ a small percentage of the U.S. population) isn't necessarily what is best for the country as a whole.
Put more directly, it's not clear that there are enough IP-based jobs to keep the United States at or near full employment. To do that, which we need for social stability, because high unemployment is intensely corrosive, we need to capture some of those manufacturing jobs. And it might be worth doing things that hurt tech companies' profitability in the short term -- taxing outsourced manufacturing, for instance -- in order to achieve that.
If we don't, the "rational business climate" might drive all the manufacturing overseas, leaving the U.S. with a very small number of employed people doing the design work, and hordes of disaffected, unemployable people with nothing to do. As you point out, once you have the production line set up in China, it makes sense to just scale it out there rather than build a second one here.
The long-term social effects of such a 90/10 split would be very bad, and would in all likelihood eventually make it very hard to be very profitable at all. So it might make sense for IP-centric design companies to take a short-term hit and try to maintain something closer to full employment, lest they find themselves first up against the wall when the welfare proletariat they created revolts.
In that case, it is likely that the skilled will hire the unskilled as servants, making the skilled people even more productive (every hour spent doing laundry is an hour spent not designing the latest iPhone). People will probably also work fewer hours, and take productivity gains in the form of leisure rather than money.
You're right on that since we're clearly talking past each other :-) Yes, businesses become commodities over time. We agree here. However this is orthogonal to innovation and job creation. If an old mature industry is on its last throes, then by all means, farm that out. Those jobs are gone. But that's not what we're talking about here. We're talking about new industries that if they can stay put for even say 5-10 years could provide immense benefits to the innovating nation. It keeps the middle class healthy and prevents the have/have-not scenario that you mentioned as troublesome. I share those same concerns. It also creates an in-house knowledge base for future innovation and an investing base for venture capital formation (think Paypal/Google mafia). It also has the effect of incenting the country to implement pro-innovation policies to continually lay those golden eggs. A virtuous cycle if I ever saw one.
Now enter the bureaucrats in Beijing who have created their own "incentives". The main one being cost which they obtain by mortgaging their labor/environment and toying with the exchange rate. These aren't truly economic advantages but are really artificial rules put in place by a planning committee for national interest. Government intervention if you will. Entrepreneurs and startups who must factor in cost to any decision are thus improperly incented to quickly offshore. Thus we have the familiar story of "we innovate and they get the jobs" This is why Grove said creating more startups won't do anything to help US unemployment since they do their scaling elsewhere. The proper course of action is for China to due away with their policies but we can't make them. This is game theory folks. The unfortunate reality is that the rules of engagement are written by bureaucrats and Grove is saying if China is playing that game we must too.
http://en.wikipedia.org/wiki/Suez_Crisis#Financial_pressure
It is easy to be protectionist. But a major selloff of the USD by China would create the mother of all confidence crises, and, while _everyone_ would lose, the world's reserve currency, the USD, would be forever damaged.
This is not a war you can realistically fight to win. I really expected more from someone of Grove's intellect. Much much more.
Is it really healthy for Silicon Valley if Apple has a factory cranking out iPhones in Cupertino, Google has a 99%-identical factory for Droids in Mountain View, and Palm has yet another factory producing Pres in Sunnyvale? What's the point? The hardware's all very similar from a production viewpoint, or it soon will be. In a rational business climate, all of these products would be made on the same line, by a company that sells manufacturing capability to the highest bidder. Competition will happen on a purely IP-based playing field, not on the factory floor.
Going back to the Tektronix example, when Tektronix finished one of those CRTs, they practically had a one-of-a-kind relic on their hands, a part that would not only not fit in any one else's oscilloscope, but could in most cases be used only in a single model in their own product line. They didn't have what we'd call excess capacity, and if they did, they darned sure wouldn't use it selling excess tubes to the competition, because those tubes were a big part of what made their overall product competitive. So every manufacturer of those particular widgets had to either build their own CRTs or buy them from a common vendor (in which case their scopes sucked).
Tektronix in 1960 couldn't afford to treat production as a red-headed stepchild in their overall business plan... but they sure can today, when the bulk of the test equipment they sell is OEM'ed out of China just like everything else. Sorry, but IMHO, the production of solar panels, portable fusion reactors, or whatever is just going to take the same inevitable course. The business will be commoditized if it prospers at all. Trade policies and externalities have absolutely nothing to do with that.