The risks mentioned in the comments above around employee retention and IP protection are very real, but in the end the risk of not being in China is far greater.
The needs of local Chinese customers (and other emerging markets) are simply different than what is required in most of the western world today. Product lines based in China will have the best ability to translate those needs into device specs and execute on them.
Here's what I've heard lately from a colleague of mine in Beijing:
He wrote to me:
"Actually, the job market of chip designers in China are still very strong. And engineers are easy to find another good salary job if they do not like the current salary, corporate culture or even the logo or slogan of a company."
I do not believe in miracles, either. No doubt a new chip taped out must be a spin-off product...but then, what we can't lose sight of is that these spin-offs must come to the market, very quickly. Otherwise, you lose the window of opportunity in meeting particular needs of a particular customer. Again, we can discuss who is more experienced in or better at meeting the spec as much as we want, but if we don't have a product designed to meet the spec in the first place, we are not even participating in the market opportunity.
Thus, I think TI is making a great move here by installing its MCU design center in Shanghai.
In my experience the emphasis in China is on cost and speed. Quality, meeting specifications, test coverage, etc., not so much. And, keep in mind that it is pretty easy to tape out a spinoff product in 6 months, which is based on a family which took TI years to develop. Furthermore, a lot of those "quick" products developed in China by Chinese companies are knock-off second sources of US IC-s. While Chinese engineers are willing to work harder and longer hours than their US colleagues they have much less experience. In general I do not believe in miracles, like developing a significant new IC from scratch in 6 months.
Replay available now: A handful of emerging network technologies are competing to be the preferred wide-area connection for the Internet of Things. All claim lower costs and power use than cellular but none have wide deployment yet. Listen in as proponents of leading contenders make their case to be the metro or national IoT network of the future. Rick Merritt, EE Times Silicon Valley Bureau Chief, moderators this discussion. Join in and ask his guests questions.