After a video reaches a certain number of views, Ted Hamilton, a product manager for YouTube analytics, explained, YouTube tells the database to freeze the view count until YouTube can manually verify the correct count to protect against botting attempts — i.e. using automated computer processes to artificially inflate the number of views.

YouTube view counts are initially tracked by servers near the end user. By looking at reports from these individual servers, YouTube engineers can detect suspicious patterns in the data.
Shared publiclyView activity