While remotely plausible, this sounds farfetched to me, considering that there would be virtually no R&D budget to deal with, and unlike chemical pharmaceutical

While remotely plausible, this sounds farfetched to me, considering that there would be virtually no R&D budget to deal with, and unlike chemical pharmaceuticals it is easy to test the full life cycle of the device in seconds, thousands of times in a row. Epinephrine is in the public domain, and there is no chance that flaws in the delivery device would take years to discover. So in comparison to most of what Big Pharma spends money trying to bring to market, an epi pen is (in comparison) extremely easy. My guess is that even if a patent were successfully filed for a new delivery device, the challenge lies in getting all the sweetheart deals that Mylan was able to get.