A quotation is not a price. It is a position on a volume curve. Ask a fabricator for “the cost of this part” and any number you get back is a guess dressed as a number, because the dominant cost in most metal parts is either a one-time tooling charge or a per-part labour and material charge, and the balance between them flips as quantity rises.
This page sets out the arithmetic of that flip, so you can price your own programme instead of collecting quotes and hoping. The formula is exact. The inputs are your own quotations, not market averages we invented.
Why a single quote tells you almost nothing
Every metal part carries three kinds of cost, and they behave differently as volume grows.
- One-time cost — tooling, die design, pattern or fixture, first-article inspection and programming. This is paid once and does not change with quantity. Spread over 100 parts it is enormous; spread over 100,000 it is invisible.
- Per-part cost — material, machine cycle time, labour, finishing and inspection. This falls slowly with volume, mainly because setup time is amortised and material buying gets better, then it flattens out.
- Volume penalty — the costs that appear only when you make a lot: inventory carrying cost, a larger exposure if the batch is wrong, and the cost of a change of mind. This one rises with quantity.
The practical consequence: processes with no tooling are cheap at low volume and expensive at high volume. Processes with tooling are the reverse. Optimum cost is almost never at the smallest quantity, and it is not always at the largest either.
The break-even formula
For any two routes to the same part, the quantity at which the tooled route overtakes the untooled route is:
Q* = tooling cost ÷ (untooled cost per part − tooled cost per part)
That is the whole calculation. Read the table below by finding your tooling cost on the left and your per-part saving across the top; the cell is the quantity at which the tooling has paid for itself.
| One-time tooling cost | Save 0.20/part | Save 0.50/part | Save 1.00/part | Save 2.00/part | Save 5.00/part | Save 10.00/part |
|---|---|---|---|---|---|---|
| 500 | 2,500 | 1,000 | 500 | 250 | 100 | 50 |
| 1,000 | 5,000 | 2,000 | 1,000 | 500 | 200 | 100 |
| 2,000 | 10,000 | 4,000 | 2,000 | 1,000 | 400 | 200 |
| 4,000 | 20,000 | 8,000 | 4,000 | 2,000 | 800 | 400 |
| 8,000 | 40,000 | 16,000 | 8,000 | 4,000 | 1,600 | 800 |
| 15,000 | 75,000 | 30,000 | 15,000 | 7,500 | 3,000 | 1,500 |
| 30,000 | 150,000 | 60,000 | 30,000 | 15,000 | 6,000 | 3,000 |
| 50,000 | 250,000 | 100,000 | 50,000 | 25,000 | 10,000 | 5,000 |
Two things about this table are worth internalising before you use it. First, the break-even moves in a straight line with tooling cost and inversely with the saving, so a tool that costs twice as much needs exactly twice the volume. Second, and more dangerous: if your per-part saving is overstated by 25%, the break-even quantity is understated by 20%. A tool at 4,000 that you believe saves 1.20 per part breaks even at 3,333 parts; if the real saving is 0.90, it breaks even at 4,444. Always run your break-even with a pessimistic saving, not an optimistic one.
Which processes win in which quantity band
Processes differ mainly in how much one-time cost they carry. That is what decides which band they win in.
| Quantity band | Usually the cheapest route | Why |
|---|---|---|
| 1 to 50 | Laser cutting, CNC bending, welded fabrication, CNC machining | No tooling at all. Every unit carries its own setup, but there is no charge to recover. |
| 50 to 500 | Laser cut and formed, or simple tooling if the part is developable | A simple or compound tool may already pay for itself, but only where the geometry is stable and the annual quantity is credible. |
| 500 to 5,000 | Compound or single-station stamping; deep drawing for axisymmetric parts | Tooling amortises across the band and per-part cycle time starts to dominate the quote. |
| 5,000 to 50,000 | Progressive die stamping; deep drawing with transfer | Multiple operations run in one press stroke, so labour per part collapses. Material becomes the largest line item. |
| Above 50,000 | Progressive dies with automated feed | Per-part cost is now almost entirely material utilisation and scrap. Nesting and strip layout decide the price. |
Note what this table is not: it is not a price list, and the bands are not walls. A part with one tight bore can justify machining at 20,000 units, and a part with a stable geometry can justify a compound tool at 300. The bands describe where the economics usually sit, and the break-even formula above is how you test your specific case.
How much does custom metal stamping cost?
The honest answer is that stamping cost is dominated by two things you control before you send the drawing: the number of stations in the die, and how much scrap the strip layout produces.
A single-station tool is the cheapest tool and the slowest route. A compound tool cuts and forms in one stroke. A progressive die runs several stations in sequence and carries the highest tooling cost and the lowest per-part cost — which is exactly why the break-even table above matters more for stamping than for any other process here.
Inside one process, the per-part cost is moved by strip layout (how much material leaves as scrap), press tonnage (which machine the job occupies), cycle time, and whether secondary operations such as tapping, deburring or plating are needed. Change one of those and the quote changes even though the drawing did not. That is why two fabricators can both be right and still be 40% apart.
What we can tell you without a drawing: we run 25 to 400 ton presses with progressive, compound and single-station dies, and we design, build and maintain the tooling in-house. Minimum order is from 1 piece, so a prototype or a bridge batch does not require a die at all — laser cut and formed parts are the cheapest way to prove geometry before tooling is committed.
How much does deep drawing cost?
Deep drawing is the clearest example of a process whose cost curve bends sharply. Tooling is unavoidable, and for parts needing several draws the tooling is a sequence rather than a single die, so the one-time cost is higher than simple stamping. Once the tool exists, per-part cost is low and stable, and material utilisation is unusually good because a drawn part uses a circular blank with little scrap.
That combination makes deep drawing a poor choice for 50 parts and a very good one at 5,000. The other cost lever is the draw ratio: asking for a deeper draw in one operation, or a tighter wall tolerance, adds tooling stages, and each added stage pushes the break-even quantity up. If the depth is negotiable and the quantity is not yet certain, that is the first thing to discuss.
How much does laser cutting cost?
Laser cutting has no tooling, so it wins at low volume and loses at high volume. Cost is driven by cut length, material thickness, the number of pierces, and nesting efficiency.
Two practical points. First, cut length is what you pay for, so a design with many small holes and intricate profiles costs more than its outline suggests. Second, at 12,000 W the machine cuts thick plate faster, but power does not help thin gauge much — the limit there is positioning speed and pierce count. We cut carbon steel up to 25 mm and hold around ±0.05 mm on profile, which varies with material and thickness.
Where laser cutting is genuinely the right answer is the band where quantity is real but not yet certain: cut and form 50 parts, learn something from them, then commit to tooling. Since single pieces can be cut, the cost of that decision is small.
How much does 5-axis CNC machining cost?
Machining has no dedicated tooling, so its cost is machine time plus programming, and it is usually the most expensive per unit of the processes here. It is also the route that holds the tightest tolerances and the most complex geometry.
The cost lever that matters most is not the machine, it is the feature count that needs tight tolerance. A tolerance like ±0.005 mm is achievable, and we run a 5-axis machining centre for that class of work, but it dictates tooling, feeds, coolant strategy and inspection time. On most parts only one or two features genuinely need it. Nominating those, rather than applying one tight tolerance across the drawing, is usually the single largest saving available on a machined part. Machining is also frequently cheapest as a secondary operation on a part that was already cut and formed, rather than as a way to make the whole part.
Why the first batch is the most expensive batch
Three costs land on batch one and never come back: the learning curve on the shop floor, the first-article inspection, and the risk that a change is needed after the die is cut.
Batch one is not the batch to squeeze. It is the batch to inspect properly, because the cost of a first article is small next to the cost of discovering at batch five that a die needs rework and 4,000 parts are out of tolerance. If you want a lever on first-batch cost, pull it on quantity rather than on inspection: make batch one smaller and inspect it fully, then release the volume.
How order quantity changes lead time
Lead time is not proportional to quantity, and the reason is setup. A single-piece order is dominated by programming, material ordering, setup, queueing and inspection, so a large part of its lead time is fixed. As quantity rises, the fixed share is amortised over more units and lead time grows much more slowly than volume — until a point where a new constraint appears, usually material availability, a shared press slot, or a finishing line that has to batch your parts with someone else’s.
Two consequences for planning. Ask for the lead time of the batch size you will actually order, not the one you were quoted as a test. And if the schedule is tight, the highest-value thing you can do is send a complete drawing package, because every clarification round adds days and every day added at the front delays release at the back.
How to compare two quotes properly
Quotes that look comparable often are not. Before you compare the per-part number, confirm that both are for the same scope:
- The same drawing revision, stated on the quote.
- The same material and temper, from the same grade, and whether the material is fixed at order or priced at delivery.
- The same finish and specification, and who inspects it.
- The same inspection scope — a full dimensional report and a visual check are very different costs.
- The same Incoterm — ex-works and delivered-duty-paid are not the same number, and freight per unit falls steeply as quantity rises.
- The same packaging, and whether tooling is charged once or amortised into the unit price.
- Whether tooling is owned by you or retained by the supplier, and what happens to it if you move.
A quote that amortises tooling into the unit price hides the break-even, and it makes switching suppliers later expensive. Prefer a quote that shows tooling as a separate one-time line.
Seven ways to reduce cost per part without changing suppliers
- Nominate the one or two dimensions that are functionally critical instead of applying tight tolerances across the drawing.
- Let the supplier influence the flat pattern and the strip layout; material utilisation is often a larger cost line than machine time.
- Standardise the material grade across parts so the shop can buy in larger lots.
- Move deburring, tapping and marking into the die rather than into a second operation.
- Batch the finishing across parts where the specification allows it.
- Ask for the break-even quantity rather than a single price, so you can see whether buying more is actually cheaper.
- Send one complete package the first time. Clarification rounds are paid for in delay, and delay is a real cost.
What these numbers are, and what we deliberately do not publish
The arithmetic on this page is exact and you can reproduce it: break-even quantity is tooling cost divided by per-part saving, and the table is that division applied to a grid of inputs. Nothing in it depends on our opinions.
What we do not publish is a price list, and the reason is not evasion. A per-part price without your drawing, material, tolerance class, finish, inspection scope and quantity would be a number that changes the moment we see the part, and a published list would mislead more buyers than it helped. What we will do instead is give you the two inputs the formula needs — a tooling figure and a per-part figure for each route worth considering — and let you run the comparison yourself.
For our own verified parameters, the quality systems we hold and the tolerances we will commit to in writing, see how to vet a metal fabricator, and for a worked comparison of a factory against an on-demand platform see XCWY against Xometry and Protolabs.
What we need in order to give you a real number
Send a 3D model in STEP or IGES, or a controlled 2D drawing with its revision level, plus material and temper, quantity, surface finish, the dimensions that are functionally critical, and your inspection expectation. We accept PDF, JPG, PNG, DXF, DWG, STEP, IGES and ZIP packages up to 20 MB, and an NDA can be signed before you send anything. If the package is complete we target a quotation within 3 business hours, and for a cost question we will give you the break-even quantity as well as the unit price, because that is usually the number you actually need.
Start from the process closest to your part: metal stamping, deep drawing, fiber laser cutting, CNC bending, 5-axis CNC machining, welding and assembly, surface finishing, or sheet metal fabrication.
Frequently asked questions
Related Reading
How much does custom metal stamping cost?
The cost of a stamped part is dominated by the number of stations in the die and by how much material the strip layout wastes. A single-station tool is the cheapest tool and the slowest route; a progressive die costs the most in tooling and the least per part. Because tooling is a one-time charge, ask for the break-even quantity as well as the unit price: break-even equals tooling cost divided by the per-part saving against the untooled route.
What is the break-even quantity for a stamping or drawing die?
Break-even quantity equals the one-time tooling cost divided by the per-part saving. A die costing 4,000 that saves 1.20 per part breaks even at 3,333 parts. Because the relationship is linear, a tool costing twice as much needs exactly twice the volume, and if the per-part saving is overstated by 25% the break-even quantity is understated by 20%.
Which process is cheapest at low volume?
Below roughly 50 parts, the processes without dedicated tooling win: fiber laser cutting, CNC press-brake forming, welded fabrication and CNC machining. Each part carries its own setup, but there is no tooling charge to recover. Laser cutting and bending in particular can prove a design before any die is committed.
At what quantity does a progressive die make sense?
Progressive dies usually start to beat single-station and compound tooling somewhere between 5,000 and 50,000 parts, because several operations run in one press stroke and labour per part collapses. The exact figure depends on your tooling quote and the per-part saving, so run both through the break-even formula rather than assuming a threshold.
How much does deep drawing cost?
Deep drawing always needs tooling, and where the part needs several draws the tooling is a sequence rather than one die, so the one-time cost is higher than simple stamping. Per-part cost is then low and stable, and material utilisation is good because a drawn part uses a circular blank with little scrap. Asking for a deeper draw in one operation or a tighter wall tolerance adds tooling stages and pushes the break-even quantity up.
How much does laser cutting cost?
Laser cutting has no tooling, so it wins at low volume. Cost is driven by total cut length, material thickness, the number of pierces and nesting efficiency, which is why a design with many small holes and intricate profiles costs more than its outline suggests. We cut carbon steel up to 25 mm and hold around ±0.05 mm on profile, varying with material and thickness.
Why is 5-axis CNC machining more expensive per part?
Machining has no dedicated tooling, so its cost is machine time plus programming, and it is usually the most expensive per unit of the processes here. The largest available saving is to nominate the one or two features that genuinely need a tight tolerance instead of applying one tight tolerance across the drawing, because a tolerance like ±0.005 mm dictates tooling, feeds, coolant strategy and inspection time.
How does order quantity affect lead time?
Lead time is not proportional to quantity, because a large share of it is fixed: programming, material ordering, setup, queueing and inspection. As quantity rises that fixed share is amortised and lead time grows much more slowly than volume, until a new constraint appears, usually material availability, a shared press slot or a finishing line that batches your parts with other work.
What is the minimum order quantity for custom metal parts?
At XCWY the minimum order starts at 1 piece, and single pieces can be cut, formed or machined for prototypes and samples. Where a process needs dedicated tooling, such as stamping or deep drawing, the practical first step is a first article from the die, so wall condition and critical dimensions can be approved before a volume run.
How should I compare two fabrication quotes fairly?
Confirm both quotes cover the same drawing revision, the same material and temper, the same finish and specification, the same inspection scope, the same Incoterm and the same packaging. Then check whether tooling is a separate one-time line or amortised into the unit price: an amortised tooling charge hides the break-even quantity and makes switching suppliers later expensive.
Leave a Reply