If you are buying a precision tool that has never before existed, AND you are pushing the technology in a number of areas ( materials science, dimensional tolerancing, production rates), then the bidding process becomes more Art than Science. The closest analogy would be the military contracting process. And we know that cost overruns abound in that arena.
If a purchasing manager neglects that reality he soon runs himself out of good suppliers. I've been involved with the toolmaking business for 25 years, and the better purchasing managers are the ones with some kind of technical backround. The MBA-types are usually more adept a blame shifting.
Most military cost overruns are a product of procurement stretchouts and requirements changes than due to real trouble hitting known performance targets.
Bottom line: if you're demanding that the customer spend a s**tload of money to get your widget to work after he's already spent a s**tload of money on your product, don't expect him to be enthused about your product, no matter how good it eventually is after his in-house staff finishes fixing it.