Cheap at volume — if the line is there
The same part, with annual demand an order of magnitude apart, wants a different process.
At the small-batch stage the mould is soft and the work is getting shape. Fit-up and the vehicle interface right by hand — labour is a large share of the piece price, and that tuition cannot be skipped. As volume arrives, the mould goes to steel and the part goes onto HP-RTM: preform cut automatically and laid robotically, with injection weight. Mould temperature and cure time held by the machine, so the only labour inside the line is load, unload, inspect. That is where the cost curve turns down.
So the move that matters at volume is changing the line, not squeezing the supplier. Leave the line where it is and every point taken off the price comes back through quality.
Cheap at volume comes in two kinds. Pultrusion is priced by the metre and gets cheaper as the run lengthens: what it saves is material and mould — near-net shape, little scrap, one die per section. HP-RTM saves labour and spreads the mould: 3–5 min a shot. Both are friends of volume; they save on different items in the quote.
What will not come down at volume is hand lay-up and vacuum bagging: the labour is all in the lay-up. Volume spreads the mould but not the hours. Winding sits in another bracket — the machine winds, people load and unload the mandrel; it comes down with volume, but not to HP-RTM levels. Those parts run on a different pricing logic; do not apply HP-RTM arithmetic to them.
Three more things are just as unavoidable in series production: mould life, batch consistency, and the tolerance stack that lets the part fit the vehicle. Mould life sets the amortisation, consistency sets the scrap rate, the stack decides whether it goes on the car. All three go into the quote; none of them can be added after series start.