I agree it is good to be product-based company. It certainly scales better than a service-based company.
However, one distinction that my Rambus interviewee was making, I believe, was between being a technology licensor (that teaches customers how to fish) rather than a patent licensee (that warns people that if they are thinking about designing a fishing rod they have to pay up).
I too believe it is always good to be a product based company rather than doing just IP licensing if the product is right for the market. Again it is encouraging to see that RamBus is working on some product capable of IoT. This trend emphasizes that IoT is on the horizon, even though there are some speculations raised about its success in the article "Internet of Things need glue".
Rambus has signed some very big deals and achieved some very big settlements over the years.
In general it's patent licensing play would seem to have been successful BUT it is hard to build a relationship with customers through the confrontational legal system AND such settlements are "lumpy" which can be frightening for shareholders.
I think Ron Black has decided that it is better to be collaborative and take a little bit less out of the customers at each licensing bite but win repeat business, just like ARM.
But he also has to decide in which markets to play as a product or service provider and in which markets to be an IP licensor. You can't do both in the same market at risk of competing with your customers.
Good move indeed. How ever, emulating an ARM-like business model will take changes to both R&D and marketing for Rambus. Licensing patents vs. Silicon-verified IP does take more R&D investment for the latter.
At a recent meeting lunch meeting of licensing executives in Silicon Valley, Rambus was listed as one of the top five non-practicing entities by revenue, so they are definitely seen as a patent, not a product company.