I guess the same thing will happen that happened to big data.
As it turned out data is not that big after all and small teams with a handful of people can tame even datalakes and enormous amount of events in realtime by now.
The point is: only a very specialized set of problems will warrant huge teams and scientific research.
