Die Casting vs CNC Machining: Where the Break-Even Actually Sits

A die-cast housing and a machined housing can look identical on a bench. The difference is where the money sits. Die casting front-loads cost into a steel tool and then produces parts cheaply. CNC machining has almost no tooling and carries its cost in every part. Choosing between them is therefore not a question about quality or appearance, but about how many parts you will buy and how much of the geometry can survive being cast rather than cut. The comparison of die casting vs CNC machining below is built around the numbers that decide a quotation, not the ones that fill a brochure.

This article sets out the real trade-off: what each process can hold, what drives cost in each, and how to compute the volume at which one overtakes the other using your own numbers rather than a generic rule.

The structural difference

Die casting injects molten metal into a hardened steel mould under high pressure. The tool is expensive and takes weeks to build, but once it exists, a part can be produced in seconds with very little labor. The process is used mainly for aluminum, zinc and magnesium, and it requires draft angles, uniform walls and generous radii so the casting can be ejected and so the metal fills the cavity before it solidifies.

CNC machining cuts the shape from solid stock. There is no tool beyond fixtures, so the first part and the ten-thousandth part are made the same way, and design changes cost a program edit rather than a new mould. The trade-off is time per part and the waste of material removed as chips.

The consequence is simple: die casting wins on unit cost at volume and loses on flexibility, precision and setup time.

Die casting vs CNC machining: where they actually compare

Two of these rows decide most projects. Tolerance and design change cost. If the part is a bracket with clean geometry and a ±0.3 mm tolerance, casting is viable. If the part is a manifold with sealing faces that must hold ±0.05 mm and a bore that must be round within 0.02 mm, it will be cast and then machined on the critical features regardless of volume, which changes the cost model entirely.

The break-even calculation

The comparison reduces to two lines:

• Cast part total cost = tooling ÷ quantity + unit cast price + machining of critical features
• Machined part total cost = unit machined price (no tooling)

Set them equal and solve for quantity:

Break-even quantity = tooling cost ÷ (machined unit price − cast unit price)

An illustrative example, using round numbers only to show the mechanic: suppose a machined housing costs 42 per part, the cast part costs 14 per part, and the tool costs 18,000. Break-even is 18,000 ÷ (42 − 14) = 643 parts. Below that, machining is cheaper in total. Above it, casting pays back, and the payback accelerates with every additional part.

Three things usually break that clean arithmetic:

1. Post-casting machining. Sealing faces, bearing bores and threaded holes are normally machined after casting. That adds a second supplier or a second operation, plus fixtures and handling.
2. Tool life and maintenance. A die has a finite life, and repairs or a replacement tool reset the amortisation mid-program.
3. Scrap and rework. Porosity shows up in leak tests and pressure tests, sometimes only after machining has added value to the part. Casting yield is rarely 100%, and the effective unit cost must include the parts you scrap.

A realistic comparison therefore uses four numbers per process, not one: tooling, unit price, post-processing, and scrap rate. Run those four through both routes and the die casting vs CNC machining question usually answers itself within one revision.

What die casting does better than machining

• Thin, stiff, complex shells. A cast housing with ribs, bosses and mounting flanges would require a large billet and hours of machining. Casting produces the net shape in one shot.
• Integrated features. Bosses, part numbers, dates and stiffening ribs can be formed in the tool at no additional per-part cost.
• Material efficiency. A casting uses close to the mass of the finished part; a machined part may remove 70% to 90% of its billet.
• Cosmetic consistency at volume. Once the tool is proven, parts are visually identical, which matters for powder-coated or painted enclosures.

What CNC does better

• No tooling commitment. Prototypes and pilot builds can be produced without spending on a mould that may be obsolete after a design review.
• Tolerance and surface finish. Machined datums and bores hold tolerances that castings cannot, and a machined face is flat by construction.
• Material freedom. 7075, 316L, 4140, titanium and engineering plastics are all machinable; none of them are die-castable.
• Change tolerance. Revisions between builds are normal in development, and machining absorbs them at program level.
• Small quantity economics. For one to several hundred parts, machining is almost always the lower total cost.

The sequence most programs should use

The pattern that works for most hardware programs is sequential rather than either-or:

1. Prototype in CNC. Prove the design, the fits and the tolerances in the real material. Change the design as often as needed.
2. Pilot in CNC or soft tooling. Produce the first hundred to a few hundred units for field testing, certification or a launch order.
3. Convert to die casting when the design is frozen and the forecast justifies the tool. Keep machining the critical interfaces.
4. Run casting for volume, with CNC capacity as the buffer for spikes, engineering changes and spare parts.

The expensive mistake is doing this backwards: committing to a die before the design is stable, then paying for tool revisions while the product is still moving. The second most expensive mistake is staying on CNC for years after volume has justified a tool, because nobody recalculated.

Design rules that decide whether casting is even possible

If the following cannot be accommodated, the part is a machining job regardless of volume:

• Draft of 1° to 3° on all walls parallel to the parting direction.
• Uniform wall thickness, with gradual transitions rather than abrupt steps.
• Radii and fillets at internal corners instead of sharp intersections.
• No undercuts unless a slide is added to the tool, which raises tool cost and maintenance.
• Machining allowance on any face that will be cut after casting, typically 0.5 to 1.5 mm depending on size and alloy.
• Ejection and gate marks placed on non-cosmetic surfaces.

A casting that ignores these rules will be quoted with a complicated tool, a high scrap rate, or a polite refusal.

Frequently asked questions

Is die casting cheaper than CNC for 500 parts? Usually not, unless the tool is small and simple. At 500 parts the tool is still being amortised across a small quantity. The break-even for a mid-size aluminum housing commonly falls somewhere between several hundred and a few thousand parts, and the exact figure depends entirely on your tool cost and the two unit prices.

Can a die casting hold ±0.05 mm? Not on as-cast features. Cast tolerances are typically around ±0.1 mm or looser, plus draft and shrinkage variation. Achieving ±0.05 mm requires machining the feature after casting, which is standard practice for sealing faces, bores and mounting pads.

How long does a die take to build? Tool build is measured in weeks rather than days, and the first samples must then be checked and usually modified. Plan the tool as a project milestone, not as a purchase order line item.

Can I mix processes in one part? Yes, and most production parts do. A casting with machined bores and faced sealing surfaces is the normal configuration. The important thing is that one supplier controls both operations, so the machined features are located against the cast datum rather than against an assumed one.

Next step

If you are weighing a tooling investment against continuing on machined parts, send the drawing and your forecast quantity. We will quote the machined part, model the cast alternative including tooling and post-machining, and show the crossover point so the decision is made on numbers rather than on a rule of thumb.