Once an organization and its stakeholders become comfortable in a status quo, they become risk averse trying to hang on to what they have. Groups starting with almost nothing have little to lose. Changing a business model is an entrepreneurial venture. It has its risks.
For that reason, companies known for innovation start new ventures off to the side, possibly even using a different name so as to avoid muddling their brand image -- and because radically different operations don't fit either the parent company's systems or its established culture. Kyocera calls this the amoeba system (new cells split off to thrive while old ones die). It's a feature of companies famous for innovation, like 3M. Among government agencies, a similar strategy has long been given labels like "skunk works" or bootstrap projects.
Toyota evolved most of the basics of their famous production system with its supplier relationships from 1950-1965. TPS was a different kind of innovation, blending a lot of different ideas into a new operations model, which became the base for novel business models. Insiders think that Toyota could have only created TPS during that era because it was geographically close to its suppliers, but isolated from normal business thought. Not really knowing "how it is done," they experimented with physical operations, making changes that seemed to be common sense just looking at them. (Compression, Chapter 2)