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)