Part of the cost of automation being less than labor is you create unemployment. This is fine--it's how progress works--and the displaced labor creates a gap between prior cost and new cost, which eventually leads to lower prices (prices don't keep with inflation, in part because it's impossible to hold all prices at the same buying-power equivalent in an economy where the relative price of everything constantly shifts thanks to population expansion, money supply increase, and productivity gains all interacting, and the buying-power total doesn't go up until the price of something goes down).
If you create 0.1% unemployment, no big deal. That gets washed away as the workforce turns over; give it a few months to a year and there's no evidence of that blip--save that we're all a little richer. Those people got new jobs, because we're buying more stuff thanks to that 0.1% difference, and somebody needs to make and move that stuff.
If you create 40% unemployment, that's a problem. Things got cheaper... okay? Who is buying all these cheaper things? Never mind that the money hasn't moved down into the hands of the unemployed (400 times more excluded than when they were just 0.1%); the people still working just found that nearly half their work isn't needed anymore, and you know what that means: they get laid off, too.
It's economically feasible when the cost of new technology is lower, and when the risk proposition across growth spans that technology in an uptake not significantly faster than re-employment. The economy has to respond to lower costs by offering lower prices, which businesses don't do unless they're pressured (e.g. competition, desire to get 5% on 100,000,000 units instead of 15% on 10,000,000 units, etc.). The universal competition--that consumers have limited money, and that every product bought is taken from that pool, thus all products are in competition with all other products--takes some time to launder the changes and produce a fresh, new economy.
China has labor shortages and might just be in a position where the reduced labor shifts chinese labor around such that the outcome is 2% or 5% or 8% unemployment, rather than 40% or 80% unemployment. If it cuts their production costs by 20%, that's a hell of a draw for import labor--China's financial position becomes that people will trade them something (to be determined at a later date; here's money as an IOU) for something China can produce cheaper than the buyer. In the mean time, places like America suddenly see a 20% drop in the price of all kinds of goods, and can buy more; this means we need more shipping, more retail operators, more people at VISA managing your accounts, and so forth. It means poor people spend a smaller chunk of their income on clothes and cell phones (which they probably need today), and more on food (which puts demands on local refineries, chemical producers, municipal water systems, John Deere, and farmers). It means more jobs here in America.
China would get richer and America would get richer just by China reducing the cost of manufacture, at least in a projection where China doesn't create an unemployment crisis in its own country.