Trying hard not to read too much into what is clearly a small number of datapoints, but it sure looks like IPv6 adoption hit a wall on January 1. Is this an artifact of the data collection process or a real effect?
IPv6 growth during the months of January-February was pretty flat for 2015 and 2016 as well, so this doesn't seem particularly noteworthy.