Injection molding produces the cheapest plastic part per unit and the most expensive first part. CNC machining and 3D printing invert that: no tooling to pay for, but a per-part price that barely improves as quantity grows. The whole decision comes down to where those two curves cross, and that crossing point is not a fixed number. It moves with your tool cost and your part geometry, and it can sit anywhere from under a hundred parts to several thousand.
This guide gives you the formula to find your own crossover, then covers the factors that override cost entirely: tolerance, material properties, design freedom and how soon you actually need parts in hand.
The break-even formula
One equation answers the core question:
Break-even quantity = tooling cost ÷ (per-part cost of the alternative − molded per-part cost)
If a mold costs 8,000 in your currency, machined parts run 35 each and molded parts run 3 each, the saving per part is 32 and the mold pays for itself at 250 units. Beyond that point every part is pure saving. Below it you have spent money on tooling you will never recover.
Because tool cost and per-part saving are the only two variables that matter, the crossover moves a long way depending on your part. This matrix shows how far.
| 金型費 | Saving of 10 per part | Saving of 25 per part | Saving of 50 per part |
| 3,000 | 300 parts | 120 parts | 60 parts |
| 8,000 | 800 parts | 320 parts | 160 parts |
| 20,000 | 2,000 parts | 800 parts | 400 parts |
| 40,000 | 4,000 parts | 1,600 parts | 800 parts |
The pattern is worth internalising. A simple part in a cheap aluminum tool can justify molding at a couple of hundred units. A complex multi-cavity steel tool for a part that was cheap to machine anyway may need tens of thousands. Anyone who quotes a single universal break-even figure is guessing.
How the three processes actually compare
| 因子 | 射出成形 | CNC加工 | 3Dプリント |
| Upfront tooling | Aluminum or hardened steel mold required | None, fixturing only | なし |
| 大量生産時の部品単価 | Lowest by a wide margin | Falls slowly, cycle time is fixed | Nearly flat, material driven |
| Time to first part | Tooling typically 15 to 25 working days | Days | Days |
| 一般的な許容差 | Cavity tolerances to ±0.015 mm, part-to-part repeatability around ±0.08 mm | ISO 2768-C general on plastics, tighter on capable features | Around ±0.2 to ±0.5 mm depending on technology |
| 材料特性 | Full isotropic bulk properties | Full bulk properties of the stock | Anisotropic, weaker across build layers |
| Design freedom | Constrained by draft, wall uniformity and ejection | Constrained by tool access and internal corner radii | Highest, including lattices and internal channels |
| 設計変更 | Expensive once steel is cut | Free, just re-run the program | Free |
| Best volume band | Hundreds to millions | One to a few thousand | One to a few hundred |
What happens at 100, 1,000 and 10,000 parts
Around 100 parts
At this quantity a mold almost never pays for itself unless the part is genuinely expensive to machine. Additive manufacturing wins on geometry-heavy parts and machining wins where you need real material properties and a tighter fit. This is also the stage where the design is least likely to be frozen, and paying for tooling before a design settles is the single most common way to waste money on a plastic part.
If you are validating form and fit rather than final performance, our overview of SLA, SLS, FDM and MJF technologies explains which process suits which kind of prototype.
Around 1,000 parts
This is where the decision genuinely goes either way. Machined and printed parts still work, but the cumulative spend starts looking uncomfortable next to a tool that would have paid for itself. An aluminum bridge tool is usually the right answer here. It costs a fraction of a hardened steel production tool, delivers real molded parts in the correct resin, and typically covers several thousand shots before wear becomes an issue.
当社の ラピッドツーリングサービス exists specifically for this band, and it also lets you run market tests or a pilot launch while a production tool is being built.
Around 10,000 parts and above
Molding wins clearly, and the conversation shifts from whether to mold to how to cavitate the tool. A single-cavity mold makes one part per cycle and costs least to build. A multi-cavity mold costs more but slashes cost per part, and it generally starts paying off around the ten thousand mark. A family mold produces several different parts in one cycle, which suits matched sets and assemblies. Sizing cavitation to your annual volume rather than your first order is what keeps the tooling investment proportionate.
The same volume logic applies to metal parts, though the thresholds differ. Our comparison of low volume CNC machining and mass production covers where that line falls on machined components.
Cost is not always the deciding factor
Plenty of projects choose a process that costs more per part because something else matters more. Four factors regularly override the break-even maths.
材料特性
Molded and machined parts both carry the full bulk properties of the polymer. Printed parts do not. Most additive processes build in layers, so strength across the build direction is lower than strength within a layer. For a functional bracket that carries load in a known direction, this matters. For a housing that only needs to hold its shape, it usually does not. If a printed prototype passed testing but the production part must survive the same load in a different orientation, that is a real engineering risk rather than a paperwork detail.
Tolerance and repeatability
Molding is extremely repeatable once the process is dialled in, but it is not the tightest process available. Shrinkage varies by resin, typically in the region of half a percent to two percent, and it is not uniform across a part with varying wall thickness. Machining plastic gives you tighter control on individual features and no shrinkage to predict, which is why critical mating features on molded parts are sometimes machined after molding rather than molded to size.
Design freedom and constraints
Each process imposes a different discipline on the CAD model:
- Molding needs uniform walls, generally in the 0.6 to 4.5 mm range depending on resin, plus draft of around one to three degrees on faces perpendicular to the pull direction. Undercuts need slides or lifters, which add tooling cost.
- Machining needs no draft at all, but cannot produce a sharp internal corner because a round tool leaves a radius. Thin walls chatter, and deep pockets need long tools that deflect.
- Additive imposes almost no geometric limits, but support removal on internal features can be difficult and surface finish is a function of layer height.
A part designed for one process often quotes badly in another. Redesigning a machined part for molding, rather than simply sending the same model, is usually where the biggest savings are found.
Time to first part
Tooling takes weeks. Machining and printing take days. If a launch date is fixed, the honest comparison is not cost per part but total programme cost including the consequence of being late. This is exactly the case bridge tooling was invented for: run real molded parts from a fast aluminum tool while hard tooling is still being cut.
Material availability differs by process
A material that exists in one process does not automatically exist in the others, and this narrows the decision earlier than most people expect.
Molding draws on the widest resin library. Commodity grades such as ABS, PP and PE sit alongside engineering grades including PC, nylon, POM and PC-ABS blends, with filled and flame-retardant variants of most of them. Colour is compounded into the resin, so it is consistent through the part rather than applied to the surface.
Machining works from stock shapes, which favours rigid engineering plastics. POM, nylon, PC, acrylic, PEEK and PP all machine predictably in plate or rod form. Soft or rubbery materials machine poorly, and some plastics release internal stress as material is removed, which shows up as warp after the part leaves the machine. Stress-relieving the stock before finishing passes is the usual remedy.
Additive has the narrowest list. Each technology supports its own family of resins or powders, and the properties rarely match the equivalent molded grade exactly. A part printed in a nylon powder behaves differently from the same geometry molded in PA66, so treating a printed material as a direct substitute for the production resin is a mistake worth avoiding during validation.
The practical consequence is that material can decide the process outright. If the application requires a specific flame-retardant grade with a certification behind it, molding may be the only route that supplies it. If it requires PEEK in small quantities, machining from stock is usually more sensible than tooling up.
The costs that do not appear in the quote
Comparing a molded price against a machined price line by line misses several things that only show up later.
- Design freeze risk. Once steel is cut, a change to a feature can mean welding and re-cutting the tool, or replacing it. Machined and printed parts absorb changes for free. If the design is still moving, that flexibility has real monetary value.
- Tool ownership and storage. Agree in writing who owns the mold, where it is stored, and what happens if you move suppliers. This is a contractual matter, not a technical one, and it is easiest to settle before the tool exists.
- Secondary operations. Molded parts may need trimming, tapping, ultrasonic welding, labelling or texture finishing. Machined parts often need deburring and finishing. These are per-part costs that sit outside the headline price.
- Inspection. Higher volumes mean sampling plans and documentation. That is a cost of the production stage, not the prototype stage, and it should be budgeted before the first order rather than after.
- Cash flow timing. Tooling is paid up front, before a single part ships. For a small team, a tool that pays back at 800 units is still a real constraint if the first order is 300.
A practical decision path
Rather than starting from the process, start from four questions about the part.
- Is the design frozen? If not, avoid tooling entirely until it is.
- What is the realistic annual volume, not the first order? Tooling should be sized to the annual figure.
- Which properties are non-negotiable? Isotropic strength, chemical resistance, temperature rating and regulatory compliance all narrow the material list before cost enters the picture.
- When do you need parts? A fixed launch date can justify a bridge tool that looks inefficient on a spreadsheet.
The answers usually point at one of three routes. Print for validation, then machine for pilot builds, then mold for production. Or skip printing and machine straight through if the part needs full material properties from day one. Or go directly to an aluminum tool if volume is clearly above break-even and the design is settled.
When molding is the wrong answer even at high volume
Volume alone does not justify a mold. Molding is usually the wrong route when:
- The part exceeds practical mold sizes, which for most tooling sits well inside a 1300 by 900 by 500 mm envelope.
- The material is available as stock but not as an injectable resin, which applies to several high-performance engineering plastics.
- You need many variants in small quantities, where a family of machined parts beats a family of tools.
- Wall sections are heavy and non-uniform in a way that cannot be redesigned, since thick sections cause sink and long cycles.
- The programme life is short enough that the tool will never reach its payback quantity.
Working out the right route for your part
The most reliable way to settle this is to have the same model quoted through more than one process, with your real target quantity attached. Quote a hundred, a thousand and ten thousand at once. The shape of those three numbers tells you more than any general rule, because it reflects your actual geometry rather than an average part.
That is how XY加工 handles the question. Send a CAD file with your target volume and we return pricing across 射出成形, CNC加工 そして 3Dプリント with design-for-manufacturability feedback inside 24 hours. If a mold is not the lowest-cost route at your quantity, the honest answer is more useful than the bigger order.
よくある質問
Q: At what quantity does injection molding become cheaper?
A: There is no universal number. Divide the tooling cost by the per-part saving over the alternative. Simple parts in aluminum tools can break even in the low hundreds, while complex steel tools may need several thousand units.
Q: Can injection molded parts hold the same tolerances as machined parts?
A: Not usually on individual features. Molding is highly repeatable, but resin shrinkage varies with wall thickness and material. Critical mating features are sometimes machined after molding rather than molded directly to size.
Q: Are 3D printed parts strong enough for production use?
A: For many applications yes, but printed parts are anisotropic and weaker across the build direction. Validate load cases in the actual build orientation rather than assuming the properties of the bulk polymer.
Q: What is bridge tooling and when should I use it?
A: A fast, lower cost aluminum mold that produces real molded parts while a production steel tool is built. It suits pilot runs, market tests and any project where a launch date arrives before hard tooling can.
Q: Should I machine or print my first prototypes? A: Print when you are checking form, fit and geometry, since it is fastest and cheapest. Machine when the prototype must be tested under load or in the
