This is a huge gap that filling would reap both financial and political rewards.
On the flip side, let's say you want to manufacture a new custom processor for your ipad competitor, or turbine for your new 787 competitor. You won't get those things in China, at least not without significant involvement from US companies.
Fundamentally I don't believe that. I could just be stupid.
I see the automation that is possible in the US, the infrastructure, and you tell me that its not possible for someone to make a business out of assembling things? Sure its cheap to pull 10 dextrous laborers out of the Chinese working class and make an assembly work cell out of them, what prevents a technologically sophisticated engineer from doing the same with one robot workstation and some programming? I've had boards made and assembled in the US, that is pretty straight forward, getting a tool and die shop to make cases isn't that hard either. What I haven't seen is an infrastructure that ties together all of these small shops like the ones in China, Korea, and Malaysia do.
Disruption is our middle name, why not disrupt the world of real goods?
Here is my thesis, creating a 'small' factory in the US is hindered by the inability of that factory to find work to fill its capacity. That limits its ability to make a profit, so the bids are priced to make sense when the factory is running far below its ability to produce. That makes the costs of goods it produces 'expensive' relative to a factory which is running at or near its theoretical capacity. If we had a way of keeping those small factories pipelines of work 'full', they could charge less because they would be getting more work out of their sunk costs (equipment, facilities, power, etc). In terms of comparative advantage we gain, transportation and communication benefits to the local markets.
So tell me why that isn't possible?
Your thesis is wrong not because it can't be done, but because the competitive gains (transportation, communication) don't outweigh the gains to doing much higher value-added tasks. (Not to speak of the potential costs in having to pay higher wages, deal with environmental concerns, etc, but i'll leave those out because they're political, not fundamental.)
Let's say the US only has 10 engineers, but they're really really good. China has 50 engineers, but they're less good. (Good being a stupid shortcut to saying they have more domain expertise/resources, not Jingoism about the quality of Chinese engineers).
Now the US engineers _could_ start a factory making iPods, and they'd be just as good as the Chinese, or just a little better and sell for around the same price, with a profit $X. It's relatively low startup cost, and has relatively low barriers to entry.
Or, the US engineers could start a factory making specialized composite-material airframes for space ships, which the Chinese lack the ability to build at all (so far), with a profit of $50 * X. It's much higher startup cost, and has huge barriers to entry, so there won't be competition for years.
The US engineers can build the wings, buy iPods from the Chinese engineers, and still pocket a huge difference! Meanwhile the Chinese engineers are thrilled to build ipods, because it's the first step in learning the skills to build Rockets. (Nonsense example, but you get the idea) This is basic comparative advantage with respect to a value-added chain.
Your theory would override this IF we had unlimited engineers and a basically level playing field for production costs at the low end, but we don't for tons of reasons, mostly around regulation (which I'm not saying is a blanket bad thing).
With only so much capital, why make iPods when you could be making satellites?
Now of course this example is trivial for lots of reasons, but the general pattern has been playing out for 100 years.
As nations gain expertise and working capital, they move up the value chain and outsource what they lack the resources or environment to build cheaply. Historically this gain in ability has been accompanied by an associated gain in wages/worker protections (for lots of complex reasons), which even further propels you a nation to move up the value chain, since you only want to build the most profitable possible things that you can.
I am not an economist or a business guy - I move bits for a living. But I'll take a stab at it.
what prevents a technologically sophisticated engineer from doing the same with one robot workstation and some programming?
I think that, right now, there ain't no such automation that can deliver what you want at a price/time you'll be willing to live with.
I've seen demos of such. Give it 10 years and you'll be able to custom order anything you can design, and that is legal.
Quasi-legal items ... well that's what your basement is for.