I'd be surprised if something similar to python 2/3 happens. The go team have been very explicit in saying that all go 1 code must continue to compile, and transitioning to go 2 needs to be as seamless as possible (most likely using tooling to automatically migrate code across, a la go fix from the early days).
Google had Guido working there for a very long time; there's probably a lot of institutional memory built up around the 2->3 transition, and likely a lot of lessons learned, and a strong desire not to repeat the experience.