How CNC Machining Price Per Part Falls From 1 to 1,000
Every buyer eventually notices the same thing: the second quote for 100 pieces comes back at a quarter of the single-unit price, and the quote for 1,000 pieces is barely lower than the one for 100. The curve is not linear, it is not a discount, and it is not arbitrary. It is the arithmetic of a fixed block of preparation spread over a growing number of parts, with two much smaller steps layered on top from material buying and fixture design. Once you can sketch that curve for your own part, you can tell whether a quoted price is reasonable, and which quantity break is worth chasing.
The curve has three distinct regions
From 1 to about 10 pieces, the price is dominated by preparation. Programming, setup and a soft-jaw fixture might total $420 on a part with 40 seconds of actual cutting per feature. At one piece, that $420 sits entirely on one unit. At ten, it is $42 each. The curve is nearly vertical and then bends hard.
From 10 to roughly 100 pieces, the curve flattens into a gentle slope. What is left is material buying, a modest amount of run time and per-piece finishing, all of which improve only slightly with volume.
Above 100 pieces, the price approaches the true marginal cost of production — material plus cycle time plus consumables — and additional quantity buys very little. The difference between 500 and 1,000 pieces on a typical machined housing is often 4 to 6 percent.
The practical consequence: if you are at 40 pieces and wondering whether to jump to 200, the answer depends on whether a fixture redesign happens in between, not on the quantity itself.
Setup amortization, the first steep drop
A three-axis part with two setups needs 4 to 5 hours of CAM and 1.5 to 2.5 hours of machine setup for the first article. Even at a modest $45 per hour for programming and $60 for machine time, that is $300 to $430 of work that exists whether you order one part or one hundred.
Divide it by quantity and the effect is stark. At one piece the fixed block adds $420 to the unit price. At ten it adds $42. At fifty, $8.40. This is the entire reason a $590 prototype becomes a $160 part at 50 pieces with no change to the drawing, the tolerance or the supplier.
Two conditions weaken the effect. First, if the part needs many setups — five-axis work on three faces, or a weldment machined after stress relief — the fixed block grows and the curve stays steep longer. Second, if the shop re-programs for each order rather than keeping your file and setup sheet, you pay the fixed block every time and never reach the flat region.
See also: What Drives CNC Prototype Cost, Line by Line and The Real Cost of Holding Machined Parts in Inventory.
Material quantity breaks are smaller than buyers expect
Ordering 500 pieces instead of 50 does not cut aluminum cost per part by 40 percent. Mill pricing for 6061 plate typically steps by 5 to 12 percent per quantity band, and the cut charge per piece falls from around $3 to under $1 at volume because the saw can nest parts from a full bar with less waste.
When a dedicated fixture earns its keep
Soft jaws cost $60 to $120 and hold one part per cycle with a manual load. A dedicated fixture costs $650 to $1,400 and typically holds two to six parts, cutting load/unload time and stabilising the second operation.
The break-even is straightforward. If the fixture saves 8 minutes of handling and re-datuming per part and your machine rate is $80 per hour, the saving is about $10.70 per piece. A $900 fixture pays back at roughly 85 parts. Below 60 pieces it is a loss; above 150 it is clearly correct; between those numbers it depends on whether the fixture also improves tolerance, which is often the real reason to build one. Above 150 pieces it also holds position better across lots, which usually matters more than the labour saving.
Second operations are where this matters most. A part that is machined on face one, flipped into a vise and indicated by hand loses 4 to 6 minutes per piece to re-zeroing, and the positional error from that re-datuming shows up directly in your true-position tolerance. A plate fixture with fixed stops removes both the time and the variation. This is the single most common reason a part quoted at 100 pieces looks expensive and the same part at 300 pieces looks cheap.
| Quantity | Fixed cost per part | Variable cost per part | Unit price | Index (1 pc = 100) | Drop from previous step |
|---|---|---|---|---|---|
| 1 | $420 | $170 | $590 | 100 | — |
| 5 | $84 | $168 | $252 | 43 | 57% |
| 10 | $42 | $165 | $207 | 35 | 18% |
| 25 | $17 | $158 | $175 | 30 | 15% |
| 50 | $8 | $152 | $160 | 27 | 9% |
| 100 | $4 | $148 | $152 | 26 | 5% |
| 250 | $5 | $132 | $137 | 23 | 10% |
| 500 | $2 | $128 | $130 | 22 | 5% |
| 1,000 | $1 | $124 | $125 | 21 | 4% |
The 250-piece step assumes a dedicated second-operation fixture; without it, that row would land near $150.
Reading the table against a real quote
Suppose a shop quotes $168 at 50 pieces and $159 at 100. That is consistent with a fixed block near $400 and a variable cost near $155 — a normal shape for a two-setup aluminum part. If instead you see $168 at 50 and $149 at 100, the shop is either buying material better than you assumed or is pricing to win the larger order. Either is worth asking about, politely, with the table in front of you.
Using the curve in negotiation
Do not negotiate the unit price at one quantity. Negotiate the structure. Ask what the fixed block is, ask what quantity triggers a fixture, and ask what the price would be at the fixture break-even. A shop that can answer those questions has a real cost model; one that cannot is guessing.
If your annual demand is 400 pieces used in four releases of 100, ask for a blanket order with scheduled releases. The shop builds the fixture once, amortizes it over 400, and you get the 250-piece price on 100-piece shipments. Hold the fixture at the shop, keep the program file current, and confirm in writing that the fixture is dedicated to your part number.
For a part that may never exceed 30 pieces, the lever is design, not quantity: fewer setups, looser non-functional tolerances, and one finish instead of three. Our capabilities overview shows the machine mix behind these numbers, and our insights section covers the design changes that move a part down the curve.
Frequently asked questions
At what quantity does CNC machining stop getting cheaper?
For most two-setup parts, the meaningful savings are gone by 100 to 150 pieces. From there to 1,000, expect 4 to 6 percent per doubling at best, and only if the shop changes process — a fixture, a pallet, or a different machine.
Why is the 10-piece price still so much higher than the 100-piece price?
Because the 10-piece order carries the full programming and setup block over a small denominator. At 10 pieces the fixed block is $42 per part; at 100 it is $4. That $38 gap is the bulk of the difference.
Is it cheaper to order 100 and store 90?
Usually yes, if the part will not be revised and the material is not exotic. Storage cost for 90 small parts is negligible against a 5 to 10 percent unit price reduction, though you should confirm the finish has no shelf life limit and that the packaging protects the surfaces.
Does a dedicated fixture always reduce the unit price?
No. Below roughly 60 pieces a $900 fixture does not pay back. Above 150 it usually does, and it also improves positional repeatability, which is often the better reason to build one.
Related reading
Send a drawing for a quote and DFM review.
STEP, IGES, X_T, PDF or DWG up to 50 MB. Tell us quantity and finish, and the quote comes back with tolerance and process notes, not just a price.