The problem is that state occurs right before product death. If you aren't investing that milk money somewhere new, your company is in a death spiral, regardless of how much money there is coming in today.
So if your company is only milking and not investing, then, yes, leave, immediately. Personally, I've not seen that happen often, but I have seen it.
IBM - one of the biggest R&D IT companies in the history of computing.
Microsoft - invented Windows, the XBox, Visual Studio, Silverlight, C#, the list is endless.
Apple - inventing radically new consumer products every other year.
Those examples are not what he was talking about at all.
A gigantic monolithic struts1 Asset Management product with 19k LOC in some of its JSP files. All they ever did with it was pitch it to companies, get us some minor web services to add to it and move on to the next company.
I couldn't even persuade them to use Agile or start to use Unit Testing, let alone start something new. When I suggested in some brief down time that we make something new, I was actually laughed at by the senior sales guy/project manager who was my boss.
God I hated that place so much. So glad I work somewhere that actually builds things and has a decent business strategy.
There is no hard and fast rule for when to leave a company, but the one rule that works for me is: it's time to go when it isn't fun anymore.
Stuff like this bothers me and I think paints an unfair picture of a corporate environment. Corporations aren't some Disney dichotomy of either being incredibly encouraging of change or a supervillain who purposefully punishes developers in order to line their own wallet.
The reality, I think, is that the vast majority of companies are in the middle. Change is good; change for the sake of change often goes against business principles. Why is a company shunning your 'innovation'? Is it because its unfeasible, or costly? Does it not integrate well with the product line?