I retired from Texas Instruments. During my tenure in the R&D Labs, if you did not fail at all you were just not pushing the design envelope enough. I worked on many failed products and some very successful ones. The Digital Light Processing (DLP) technology, that I worked on, took 12 years of internal R&D funding before they sold the very first chip. No other company, Chinese or not, could afford a long term project like this one. For any company to really succeed, the management must have long term goals in mind. By long term I mean many years, not just a few months.
An eight reason can be lack of IP protection. Even a reputation for lack of IP protection makes it very difficult for Chinese companies to trust each other and form meaningful partnerships. Western SoC companies nurture a third-party IP vendor (TPIPV) ecosystem based on trust. TPIPVs can dramatically reduce time to market and investment required.
It is true that culture and IP barrier handicap have been important factors, but that can be changed with money. What they need is to develop a few chips for a couple of killer applications; rake in tons of money, then they will be on their way. There are a couple of good examples in Taiwan.