How high is that. I guess the cost of SQL is less. Seems some index it should not use which it's using now and that's giving problem. I guess your CPU elapsed will be also close to total elapsed here. Look into all the indexes and check the cardinality, clustering_factor. Buffer gets SQL take s time to tune.
Did you analyze the table after adding two columns BTW?
It is not an index problem for sure. The new columns are not used in any select statement.
Very important thing is the the the SQL which is stuck is not using this new columns at all. However, the table which is modified is being used for select
(excluding the new columns) and the output in inserted into other table which not touched at all.