Hardware designers benefit from having multiple separate teams to test their product. A chip designer can rely on at least two other teams to test the designed chip, and one of them will be using formal verification. If software also has long release cycles and high cost to remedy mistakes, you bet we would also have multiple testers. In fact that was what happened in the 90s with shrink wrapped software and without easy updates.
That was the case for the company that I worked at.
The official QA organization was very powerful, and had no compunctions about stopping an entire product line, for one bug.
When that happened, the department responsible for the bug would find themselves against the wall.
As a result, all the software departments had pretty big teams of testers, who would validate the software, before it was released to the purview of the QA organization.
It could be pretty restricting, but we always felt confident that what we shipped, worked.
