EDIT: Whoa, took a while to post this and then parent was gone. I should've quoted them, ah well. Their comment was along the lines of, "Why are companies hiring people for $40k to write this stuff (and getting bad products as a result) instead of hiring someone for $140k that'll turn out better products."
Management doesn't look at it that way at these companies. The problem is hardware companies don't grok software (not just companies, US DoD has such a heavy background in hardware they often don't have a clue how to handle software, see the F-35 and attempts to "modernize" old systems, wonder why budgets on new systems get out of control? It's probably the software.). An anecdote, but one that I'm told by others was representative of their experience with hardware companies.
The company began by making sensors. It got into software because they wanted a way to sell the complete package to their customers (buy our sensors and this control panel, your building/bus/train/plane will be safer and you don't have to write the control software yourself!). Before my time there they had 4 or 5 people in software. 1 coder. 1-2 testers. 1-2 QA. Their test suites were ludicrously incomplete. No one thought that this was odd, except for those of us in software who came to the show later. The higher ups considered software to just be a necessary inconvenience, it wasn't until our shop started delaying deliveries (because it was understaffed, underpaid, and inexperienced) that they started hiring more people (all underpaid and inexperienced as well) in an effort to throw more bodies at the problem (seriously, see the F-35 for a repeat of this exact behavior).
Short version: Hardware organizations don't get software. It's an inconvenience to them, and they'll screw over the poor sap that gets stuck managing the project with a pitiful budget, and absurd deadlines and requirements.