Part of the process & material guide series — see all →
Die casting is the default answer to "make it in metal" for a lot of buyers — and the wrong one half the time. What actually decides whether a part should be cast, and how to tell before anyone spends money on a mold.
Ask a factory "can you make this part?" and the answer is always yes — every process can make almost anything, given enough machining and welding downstream. The real question is which process makes it at the cost and quality the drawing deserves. Die casting has a narrow set of conditions under which it is clearly the best answer, and a long list of situations where it quietly costs more than the alternatives. This page is about telling the two apart before tooling is cut, not after.
Die casting is the right call when three things are true at once: the part has a complex three-dimensional shape that would otherwise be assembled from several pieces or heavily machined, the volume is high enough to amortize a mold, and the alloy is one of the casting alloys — typically ADC12 or A380 — so its limits (no welding, poor anodizing, some porosity) do not conflict with the part's job. When any one of the three fails, casting usually loses to CNC machining, extrusion or sheet metal. The skill is checking all three honestly, because a quote for a casting always looks attractive until the mold cost, the machining after casting and the finishing limits are added back in.
Four capabilities justify a mold, and everything else is secondary:
Notice what is not on the list: surface finish quality, weldability, toughness and low-volume flexibility. Those are casting's weak points, and they are exactly where most wrong decisions are made.
Run the part against these five checks. All five should pass, or the exceptions should be consciously accepted:
Each of these situations has a better answer, and the earlier the better:
| Your situation | Better default | Why |
|---|---|---|
| Complex 3D housing, 10k+/yr, painted finish | Die casting | Mold amortizes; one-shot geometry; finish suits casting |
| Prototype or design not frozen | CNC from billet | No tooling committed; design freedom |
| A few hundred pieces per year | CNC or extrusion | Mold never pays back at this volume |
| Must weld to other parts | Wrought alloy route | Casting alloys cannot weld soundly |
| Clean anodized cosmetic surface | 6061/6063 extruded or machined | Casting alloys anodize poorly |
| Long constant cross-section (rail, frame, heatsink) | Extrusion + machining | Cheap die, tight tolerance, clean finish |
| Flat enclosure or bracket | Sheet metal | Fast, flat, no tooling at modest volume |
| Pressure-tight fluid part | Cast only with an early sealing plan | Porosity risk must be engineered out, not discovered |
Tooling is the pivot of the whole decision, so it deserves its own arithmetic. A casting mold is designed, cut, hardened, sampled and corrected before a single shippable part exists — think weeks of lead time and a capital cost that would buy a lot of machined parts. From there, the per-part cost is low and mostly flat. Machining has no tooling but a per-part cost that stays high forever. The two lines cross, and the crossing point is the decision. Buyers who compare quotes across processes without modeling the mold get this wrong in both directions: they cast parts that should have been machined (paying the mold off over a trickle of volume), or they machine parts that should have been cast (paying the machining premium for years).
Two practical refinements matter. First, the mold's complexity moves the break-even dramatically — a simple cover may pay out at a few thousand pieces while a multi-slider housing needs far more. Second, molds fail or need maintenance; a serious quote mentions mold life and maintenance responsibility, and a serious buyer asks. If the vendor's quote does not separate tooling from piece price, that is the first question to ask — the honest breakdown is in die casting in China: tooling, porosity and when to cast.
Most real die-cast parts are hybrids. The casting forms the complex shell; machining adds what casting cannot hold — precision bores, flat sealing faces, threaded holes, critical datums. Designing for this split is the difference between a cheap part and an expensive one: put the tight tolerances only where machining will happen anyway, give the casting generous draft everywhere else, and specify threads as cast-in inserts only where the load demands it (otherwise tap after machining — faster and cheaper). The DFM checklist covers the drawing rules; the point here is that the casting decision is not binary. Choosing casting means choosing casting plus a machining plan, and the quote should show both.
This decision — cast, machine, extrude or fabricate — is the first thing I settle on any new part, because everything downstream (tooling budget, unit cost, lead time, finish options) follows from it. I review the drawing against all four process routes with real factory numbers, not rules of thumb, and I will tell you casting is the wrong call before you commit to a mold. See how the engagement runs:
Read: die casting in China — tooling, porosity and when to cast →
Read: die-cast ADC12 vs extruded 6061/6063 — the material difference →
Send a drawing — get the cast-or-don't call before you commit →
It depends on mold complexity, but a practical range is: below roughly 1,000 pieces per year, machining from billet usually wins on total project cost; from the low thousands upward, a casting mold starts to pay for itself; at tens of thousands of pieces, casting is usually the cheapest route to a complex aluminum shape. The exact break-even comes from pricing the mold against the volume — which is why the quote should always separate tooling from piece price.
Technically yes, commercially it is usually a mistake. A mold built before the design is proven absorbs every revision as rework or a second mold. The standard route is to machine first articles from billet or plate, freeze the design, then cut the mold for volume. If cosmetic or material parity with the final casting matters for early testing, that constraint changes the calculus and should be stated up front.
This combination is the classic conflict, because the casting alloys that flow well (ADC12, A380) anodize dark and mottled. Options in order of cost: paint or powder coat the casting instead of anodizing; mask-and-machine the critical faces and accept as-cast finish elsewhere; or redesign so the cosmetic component is a wrought-alloy part (extruded or machined) joined to a cast body. Decide before tooling — this is not fixable with process tweaking.
For stiffness-dominated brackets, housings and covers, yes — casting alloys are strong and dimensionally stable. For impact or fatigue-dominated structure, be careful: casting alloys are relatively brittle (low elongation) and porosity acts as a fatigue initiator. A structural part that must flex, absorb impact or survive many load cycles usually belongs in a wrought alloy — machined, extruded or forged — or needs explicit analysis before casting is committed.
Casting is one of four in-house routes: see the die casting factory, and how a real part moved from CNC machining to die casting once volume justified the mold.
Send the drawing and target volume — you'll get the cast / machine / extrude call with the reasoning, before any tooling is committed.