I think cost reduction is just one dimension. Another dimension that is just as important is solution quality. If we could guarantee far better quality of solutions but in a shorter time then you have better solutions delivered faster resultng in a more agile enterprise. The fact that they cost less is not the driving force. The driving force is enterprise agility and higher quality (right first time as opposed to good enough at some time).
As to innovation, it rather depends on what the nature of the innovation is. If the innovation is some solution in IT that it and of itself is distributed then it enables you to test for alignment of the solution and to test for requirements satisfaction at design time which takes out a lot of cost in the proof of concept stage by moving one of the gates (satifiability against requirements) for continuation to an earlier stage. This means that you can disqualify some innovation proof of concepts faster and so waste less time and resource in doing so. Equally it means that such innovation can be done faster and with a higher quality as before.
Testable Architecture and Savara are not innovation tools any more than UML and Rational Rose are. But they do help because for those IT innovations that match what Savara can provide we can reduce the cost of the innovation as mentioned above.