One of the weak points of TRIZ is that it does not take into account that any technical system has various niches of application. Consider, for example, how this weakness manifests itself in TRIZ theory of S-curve. This theory says that when improvement of the main characteristics of a system stagnates it gets replaced by a new system. But it is reasonable to ask in what niche of application improvement of the main characteristics stagnates? It may happen that improvement stagnates in one niche but does not stagnate in others.
Generally, TRIZ knows just one scenario of how an old system gives way to a new one. It is when its improvement stagnates. But in the reality another scenario is far more frequent:
It could happen that a new system emerges in the same niche as the old system. This leads to dividing the niche in two, the one where the old system continues to reside, and the one where the new system managed to gain a foothold. As technology progresses, either niche can either shrink or expand.
It could happen that a technical system loses all its commercial niches and dies out commercially. But no technical system dies out in the niche of the research. This is the only niche where they do not struggle but co-exist. Eventually, a system that died out commercially could get improved in the niche of the research and spread to a commercial niche again.
Generally, technical systems evolution is more about their intra- and inter-niche struggle and advancement that about anything else. But this type of evolution is completely missed in TRIZ.