It actually depends on the starting point. You can find proofs for both arguments.

In a perfect world, both statements are only semantically different. Then one Data Mart is essentially one dimensional model, e.g. sales data is one, finance another. And the union of sales plus finance is the DWH.

(This statement fall short a bit already because of shared dimensions like customer dimension which would need to be in both but would be dealt with in a perfect implementation either.)

The reality however is that data marts are built by departments, with their own requirements in mind only. That's why they are so much easier to build. The sales department has sales data plus a bit of finance, the finance department similar.

The union of such is certainly not a DWH, something with the single point of truth.