"Should we cast this or machine it?" comes up on almost every new part program. The answer hinges on volume, geometry, material, tolerance requirements, and total landed cost. Here's how to work through it.
The fundamental tradeoff
Machining starts with a solid block of material and removes everything that isn't the part. It's subtractive. You pay for the raw material (most of which ends up as chips), the machine time to cut it, and tooling that rarely exceeds a few hundred dollars in end mills and inserts.
Casting starts with a mold and fills it with molten material. It's formative. You pay for the mold upfront (anywhere from $500 for a simple sand mold to $100,000+ for a multi-cavity die), but the per-part cost drops fast because you're only using the material that ends up in the finished part.
The crossover point depends on the part, but the general rule holds:
- 1 to ~500 parts: Machining is almost always cheaper. No tooling investment, no minimum order quantities.
- 500 to 5,000 parts: Gray zone. Depends on geometry, material cost, and which casting process fits.
- 5,000+ parts: Casting usually wins on cost, often by 30-50% versus machining from solid stock.
When to machine
Choose machining when:
- Volumes are low. Prototypes, short runs, and bridge production are where machining shines. A 3-axis VMC running at $75/hr can turn around a batch of 50 aluminum brackets in a few days with zero tooling lead time.
- Tolerances are tight. CNC machines hold ±0.001" or better without breaking a sweat. As-cast tolerances run ±0.010" to ±0.030" depending on process, so anything with close-fitting bores or mating surfaces will need secondary machining anyway.
- Surface finish matters. Machined surfaces hit 16 Ra or better straight off the tool. Cast surfaces are rougher and almost always need a finishing pass on functional faces.
- Material options are wide. Any machinable metal, plastic, or composite is fair game. Casting is limited to alloys with the right melting points and flow characteristics.
- Design is still changing. Updating a CNC program takes hours. Modifying a casting die takes weeks and $5,000-$20,000 in rework costs.
- Lead time is critical. Machined parts ship in 1-3 weeks. Casting tooling alone takes 8-16 weeks before you see your first article.
When to cast
Choose casting when:
- Volumes are high. A high-pressure die casting cell produces 40-120 shots per hour with cycle times of 30-90 seconds. One operator can run multiple machines. At 10,000+ annual volume, the unit economics are hard to beat.
- Geometry is complex. Internal passages, thin walls, and organic shapes that would require hours of 5-axis machining are routine in casting. Automotive transmission housings are a textbook example: complex internal ribbing, oil channels, and bearing bores all formed in a single shot.
- Material utilization matters. Machining a titanium aerospace bracket from billet typically produces a buy-to-fly ratio of 10:1 to 16:1, meaning 90-94% of the raw material becomes chips. Near-net-shape casting drops that ratio to roughly 1.5:1 to 2:1. When titanium costs $15-30/kg, that waste adds up fast.
- Weight reduction is a goal. Casting enables thin walls (down to 0.5mm in die casting) and hollow sections that would be impossible to machine from solid. Aluminum die cast transmission cases achieve roughly 40% weight savings versus iron alternatives.
- The design is locked. Once the part is frozen for production, the tooling investment pays back quickly. A $50,000 die amortized over 100,000 parts adds only $0.50 per unit.
Casting processes compared
Not all casting is the same:
| Process | Tooling Cost | Per-Part Cost | Tolerance | Best For | |---------|-------------|---------------|-----------|----------| | Sand casting | $500-$5,000 | Medium | ±0.030" | Large parts, low volume, ferrous alloys | | Investment casting | $2,000-$15,000 | Medium-High | ±0.005" | Complex geometry, tight tolerances, superalloys | | Die casting | $10,000-$100,000+ | Low | ±0.005" | High volume, aluminum/zinc, thin walls | | Permanent mold | $5,000-$25,000 | Medium | ±0.015" | Medium volume, aluminum, good mechanical properties |
Die casting tools last 150,000 to 300,000+ shots before major refurbishment, depending on part complexity and alloy. Simple brackets push toward the high end. Tight-tolerance automotive parts with slides and cores trend lower.
The hybrid approach
In practice, many parts use both processes. A common workflow:
- Cast the near-net-shape blank
- Machine critical features (bores, mating surfaces, threads, seal grooves)
Engine blocks, pump housings, and aerospace structural castings all follow this pattern. Sourcing both operations from a single supplier typically saves 20-30% compared to buying castings and machining separately. The combined cost of cast-then-machine is often 30-50% lower than full CNC machining from billet, especially for aluminum and iron parts above 1,000 annual units.
Making the call
Ask these questions in order:
- What's my annual volume? Under 500? Machine it. Over 5,000? Get casting quotes.
- What tolerances do I need on critical features? Tighter than ±0.010"? Budget for secondary machining regardless of the primary process.
- What's my timeline? Need parts in two weeks? Machine them. Can you wait 10-16 weeks for tooling? Casting becomes viable.
- Is the design finalized? Still iterating? Machine it. Frozen for production? Invest in tooling.
- What's the material? Aluminum, zinc, steel, or iron? Plenty of casting options. Titanium or Inconel? Casting is possible (and saves enormous material cost), but requires specialized foundries. Engineering plastics? Machining or injection molding, not casting.
There's no universal answer. But running through this framework gets you to the right process for your part, your volume, and your budget.