Manufacturing Path/Estimating Production Costs for Arduino-Based Project

For what it's worth, and given the implied lack of experience of commercial product development, I think it's crazy to consider jumping into a production run of thousands or tens of thousands if you don't already have substantial experience of developing something similar. You can get a dozen functional prototypes made for money which is trivial in terms of the numbers you're describing, and if you plan to scale this up to thousands I suggest you should subcontract the design and manufacture to specialists. By the time you're thinking about making tens or hundreds of thousands, you should already have sold thousands. That's my view, anyway.