Good points. I hadn't really thought through the economics of it. I wonder... if the testing could be automated at high speed, would that make the economics work? That would enable being able to service entities that buy larger quantities of parts, reducing the impact of shipping costs. Those don't have to even be companies -- I already buy most of the parts for my hobby projects in larger quantities, in order to get bulk discounts, free shipping, and to have sufficient stock on hand that I don't need to order things every time I start a new project.

Now my imagination is running wild...

I agree with this. I often get asked to decap + image chips commercially and its not usually worth it unless its a value add / loss leader as part of a larger project.