Choosing between injection molding, 3D printing and CNC machining comes down to three numbers: how many parts you actually need, how tight the tolerance really is, and how much tooling money you are willing to commit before the design is proven. Under about 100 parts, printing or machining wins because there is no tool to pay for. Above roughly 10,000 parts, injection molding spreads its tooling cost so thin that nothing else competes on unit price. The awkward middle — 100 to 10,000 parts — is where the decision genuinely depends on the part, and it is where most buyers get talked into the wrong process.
Key takeaways
- Tooling, not machine time, is what makes injection molding expensive at low volume — and cheap at high volume.
- 3D printing has no tooling and no minimum order, but per-part cost barely falls as quantity rises.
- CNC machining removes material from a solid block: the best route when a drawing needs tight tolerances or a specific engineering plastic.
- The crossover from printing to molding usually sits between 500 and 2,000 parts for a simple part, and much lower for a small one.
- The choice is not exclusive: many programmes print for validation, machine a bridge batch, then mold for production.
In this guide
- 1. The three processes at a glance
- 2. Injection molding: when volume sets the unit price
- 3. 3D printing: when there is no design to commit to yet
- 4. CNC machining: when material or tolerance rules out molding
- 5. Where the crossovers actually sit
- 6. The tooling decision nobody warns you about
- 7. The decision in four questions
1. The three processes at a glance
The table below is the fast version. Treat the ranges as indicative — they move with part size, geometry and material — but the shape of the comparison holds.
| Factor | Injection molding | 3D printing | CNC machining |
|---|---|---|---|
| Tooling required | Yes — the main cost | None | None (fixtures only) |
| Economic quantity | 10,000+ parts | 1–100 parts | 1–5,000 parts |
| Unit cost at volume | Lowest of the three | Barely falls with quantity | Falls with quantity, plateaus |
| Tolerance | Typically ±0.1 mm, tighter on critical features | Loosest of the three | Tightest of the three |
| Lead time to first part | 4–8 weeks including tooling | 1–3 days | 3–10 days |
| Material range | Very wide — production resins | Narrow — printable grades only | Wide for plastics and metals |
| Best fit | Stable design, committed volume | Concept and fit checks | Tight tolerances, low volume, metal parts |
Read down the “tooling required” row and most of the decision is already made. Injection molding is the only one of the three where you pay for a production asset before you have a single saleable part.
2. Injection molding: when volume sets the unit price
Injection molding forces molten resin into a closed steel or aluminium tool under pressure. The tool is expensive and the cycle is fast, which produces the characteristic cost curve: a large fixed cost, then a very low variable cost. That is why a mould that looks absurd at 500 parts looks cheap at 50,000.
What the process gives you beyond unit price is repeatability. Once the tool is qualified, part 50,000 is dimensionally the same as part 50, assuming the tool is maintained and the process parameters are locked. For a consumer product or a component that has to assemble with another part, that consistency is worth more than the unit saving.
What it does not give you is patience. Change the design after the tool is cut and you are paying for welding, re-machining or a new insert. Injection molding rewards buyers who have finished designing.
3. 3D printing: when there is no design to commit to yet
Additive processes build a part layer by layer, so there is no tool and no minimum order. For a design that is still moving, that is decisive: you can iterate every week without writing off tooling. Printing also handles geometry that molding cannot, such as internal lattices and deep undercuts with no witness lines.
The limits are material and surface. Printable grades cover a fraction of the resins available to a moulder, the mechanical properties are not identical to the moulded equivalent, and the surface carries layer texture unless it is post-processed. If your part has to pass a drop test at production wall thickness in a specific glass-filled nylon, a printed sample is not an answer to that question.
Used properly, printing is the cheapest way to answer fit-and-form questions before anyone commits capital. Used as a shortcut to production, it produces a part that works in the meeting and fails in the field.
4. CNC machining: when material or tolerance rules out molding
CNC machining cuts the part from a solid block. There is no tool, so there is no tooling write-off, and the achievable tolerance is the tightest of the three routes. It is also the only one of the three that covers metals and engineering plastics with the same process.
The trade is that material is bought as a solid block and most of it becomes chips, and every part costs machine time. That is why machining is economical from one part to a few thousand, and then loses to molding. It is also the standard route for parts that will never justify a tool: scientific instrumentation, fixtures, one-off replacement parts and low-volume industrial equipment.
5. Where the crossovers actually sit
Buyers often ask for a single quantity that separates the processes. There is not one, but the boundaries are narrower than most people assume.
- Printing to machining, around 10–50 parts. Printing’s cost per part is roughly flat; machining’s falls as programming and fixturing are amortised. Once you need a specific engineering plastic or a real tolerance, machining overtakes printing.
- Machining to molding, roughly 500–2,000 parts for a small simple part. A small, simple part with a modest tool can break even surprisingly early.
- Machining to molding, roughly 5,000–15,000 parts for a larger or cosmetic part. Bigger tools, texture and multi-cavity layouts push the break-even up.
- Never mold below about 300 parts unless the tool is reused for something else, or the part is so simple that the tool is genuinely inexpensive.
The honest way to answer the question is to price both routes on your actual part rather than on a rule of thumb. Break-even quantities move by a factor of five between a 20 mm bracket and a 200 mm textured housing.
6. The tooling decision nobody warns you about
The choice between the processes is usually framed as a cost question, but the more expensive mistake is committing to a tool for a design that is not finished. A tool bought too early either gets modified at cost, or locks a flaw into tens of thousands of parts.
The practical sequence that avoids both errors is: print or machine the validation parts, freeze the design, then cut the production tool. Where the market window is short, an aluminium bridge tool can supply saleable parts while a hardened steel tool is being built behind it — you pay for two tools, but you do not pay for being late.
Our answer: we make molds in-house and run trials and production through vetted partner factories, which means we quote the tooling question rather than the process we happen to own. Our toolroom runs 13 CNC machining centres with travels up to 1,600 mm, six EDM machines including Sodick mirror-finish and twin-head units, three wire EDMs and five surface grinders, so we cut critical mold dimensions to ±0.005 mm. A tool cut to that accuracy will hold ±0.01 mm on critical moulded dimensions, subject to the resin and the part geometry — two separate numbers, and worth asking any supplier to state both.
7. The decision in four questions
- How many parts will you genuinely sell in the first two years? Use a committed forecast, not a best case.
- Is the design frozen? If not, the tooling decision is premature regardless of volume.
- Which dimensions have to hold, and to what tolerance? A single tight feature can push a part to machining no matter what the volume says.
- What is the cost of being four weeks late to market? If it is high, printing and bridge tooling are cheap insurance.
Answered honestly, those four questions pick the process. Where the answer is a close call, ask for quotes on two routes against the same CAD file and compare landed cost per part, not tooling price.
For the pricing mechanics behind these numbers, see our injection molding cost guide, and for the tooling side see How Much Does a Plastic Injection Mold Cost?.
Related guides in this series
This article is part of our Injection Molding Processes Compared — a full walkthrough of the topic with the numbers and checklists behind each decision.
- Insert Molding: What It Is and When to Use It
- Insert Molding vs Overmolding: Key Differences Explained
- Rotational Molding vs Injection Molding: Which Process Is Right for You?
- Low Volume Injection Molding: A Complete Guide to Small Batch Production
Get an engineering answer, not a sales pitch
If you are working on choosing a process, send us the part and the target volume. We will come back with a design review, a tooling recommendation and an itemised quote — and if a different process or material is a better fit for your volume, we will say so. Talk to a manufacturing engineer.