I Am Not An Economist, but, the problem with infrastructure is that its difficult to know where to put it to generate a reasonable return on investment- predicting whether a new road gets used or a new residential area becomes desirable is more of an art than a science, and if you get it wrong you've essentially thrown money out the window. China has had this issue in spades- witness its infamous "ghost cities"[1]. In the long run, China's infrastructure growth (and therefore, to at least some degree, growth in general) will be limited not by what people are willing to spend on it, but peoples' understanding of the exact shape of what they're building with it.

[1] http://priceonomics.com/surveying-the-ghost-cities-of-china/