I think the sunsetting is very premature. Python 3 became a stable and viable alternative only since 3.6.1 (look at the evolution of async/await before that for example). Giving it 2 years is definitely not enough for enterprise (unless their focus is startup, tinkerers and data scientists)
Loads of open source and commercial projects have already made the switch to python 3 years ago. If all of them could do it, so could your org. No excuses, you've made your bed, etc.
