If you are following the standard QIIME 2 approach of generating a feature-table, you'll have to transpose the table, only if you are going append taxonomy, or other metadata, that is being joined on the feature-id. The transpose command does not know anything about your data, it will just transpose the table it is given. If you are unsure of which orientation your feature-table is currently in you can simply run the following to see which way the data is oriented:
qiime metadata tabulate
--m-input-file table.qza
--o-visualization table-tabulate.qzv
From this you'll see that the rows are samples, and the columns are feature-ids. But if we'd like to append other metadata, like taxonomy, to this table we need to transpose the table so that the rows are feature-ids and the columns are samples. This way we can append the taxonomy, and other metadata, which are joined to the table via the row (feature-ids).
This should match the information output from:
qiime feature-table summarize \
--i-table table.qza \
--m-sample-metadata-file sample-metadata.tsv \
--o-visualization table.qzv \
-Mike