Eitan Blumin does not like the limitations of metadata-only column changes with SQL Server 2016:
This is an excellent mechanism on the one hand…
However, it’s completely useless when the column you want to change has a CLUSTERED INDEX defined on it (regardless of whether it’s also a PRIMARY KEY or not).
Such a scenario would especially be common with IDENTITY columns (which, ironically, is exactly the kind of examples that Paul presented in his post).
Click through to understand the scope of this limitation.