Parts produced by 3D printing hit one million units. Why not use injection molding for mass production?
A production engineer's read on the point where additive stops making sense. We cover the symptoms that show up on high-volume additive runs, what causes each one, and how to decide whether injection molding for mass production is the right switch or the wrong one.
In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
When parts produced by 3D printing start failing at volume
Use the left column to find your situation. The middle column is the mechanism behind it. The right column is the practical move.
| Symptom | Likely cause | What to do |
|---|---|---|
| Unit cost stops falling after ~5,000 parts | Machine time and labor scale linearly, not with volume | Quote tooling and compare at 50,000+ units |
| Batch-to-batch color drift | No colorant masterbatch control on the print line | Move cosmetic parts to molded resin |
| Wall porosity on pressure parts | Layer bonding leaves micro voids between roads | Machine the seal face or switch process |
| Creep under sustained load | Thermoplastic layers relax over time | Re-spec material or go to a molded grade |
| Dimensional drift across a long run | No closed-loop thermal control on the bed | Add in-process checks or move to tooling |
| Operator hours per 1,000 parts stays flat | Post-processing is manual on every part | Automate finishing or tool the part |
| Scrap rate climbs after month three | Nozzle wear changes extrusion width | Requalify the process or change process |
What parts produced by 3D printing can and cannot do at volume
Additive earns its place in low-volume work. A few hundred brackets, a housing revision, a jig that only exists for one build cycle. The economics change slowly, then suddenly. Below roughly 1,000 units, machine time dominates and the print farm is fine. Between 1,000 and 20,000 units, the labor behind depowdering, support removal, and inspection starts to weigh more than the machine.
At one million units, the arithmetic is different. A printer that runs 24 hours a day makes a fixed number of parts per hour, and that number does not improve with the order size. Two hundred printers is two hundred times the labor, floor space, and maintenance. Tooling does the opposite: the cost is paid once and then spread across every part that follows.
This is why the question in the H1 is not really about quality. Additive can hold tolerance on a single part. The problem is that holding it on part number 400,000 requires the same attention as part number one. Molded parts inherit their dimensions from a steel cavity, which does not get tired.
The honest boundary: if your part is large, hollow, lattice-filled, or changes every few months, additive stays the better answer no matter the quantity. Injection molding for mass production wins when the geometry is stable, the wall sections are uniform, and the part will not change for a year or more.
- 1Stable geometrySame drawing for 12+ months, no planned revision
- 2Uniform walls2–4 mm nominal, no thick-to-thin transitions
- 3Cosmetic surfaceVisible parts where batch color consistency matters
- 4Load bearingSustained stress where layer creep is a risk
Why the unit cost curve flattens on a print farm
Print cost per part is roughly machine rate divided by parts per hour, plus material, plus post-processing. None of those three terms improve much at scale. Material is bought by the kilogram and a million-part order does not change the extrusion physics. Post-processing is per part by definition.
Tooling inverts this. A single-cavity mold costing a few thousand dollars amortized over 50,000 parts adds only cents per part. A multi-cavity tool drops that further. The break-even point moves with part size, material, and cavity count, so the only reliable answer is a quote at both volumes.
Watch the hidden terms. Additive quotes often exclude support removal, bead blasting, and dimensional inspection. Those are real labor hours. If your supplier quotes print time only, the comparison you are making is not the comparison you think you are making.
Material choice also shifts. Print resins are formulated for flow and layer adhesion. Molded grades are formulated for impact, UV stability, and creep resistance, and they cost less per kilogram in most cases. That gap widens on engineering resins like PA, POM, and PEEK.
Where additive quality actually breaks down over a long run
The first hundred parts off a well-tuned printer can be excellent. The trouble appears as hours accumulate. Nozzle wear widens the extrusion path by a few hundredths of a millimeter, which shifts wall thickness on every part that follows. Operators compensate by feel, and the compensation drifts.
Thermal history is the second issue. A part printed near the top of a tall build sees a different cooling profile than one printed at the bottom. On a 20-part build that variation is invisible. On a 200-part build it shows up as a bimodal distribution in the inspection data.
Layer bonding leaves micro voids along the Z axis. For a bracket that is fine. For anything holding pressure, those voids become leak paths, and no amount of cosmetic finishing closes them. We see this most often on manifolds and valve bodies.
Molded parts do not have this problem because the material is homogeneous. There is no layer interface, no nozzle wear term, and no build-position effect. The cavity is the same for part one and part one million.
- 1Nozzle wearExtrusion width drifts after 200–400 hours
- 2Build positionCooling profile differs top to bottom of the plate
- 3Z-axis voidsLeak paths on pressure-tight parts
- 4Manual touch laborSupport removal and inspection scale linearly
How to decide between printing and injection molding for mass production
Start with annual volume, not total volume. A part you need 30,000 of over three years is a different decision than 30,000 in one quarter, because tooling amortizes across the whole production life while print labor is paid every month.
Then look at the revision cadence. If the design changes twice a year, tooling change orders eat the savings. If it has been frozen for a year, tooling is close to free money. Engineers sometimes forget that a mold can be modified, but each modification costs a few weeks.
Next, check the tolerance stack. Additive holds ±0.1 mm on a good day. If your assembly needs ±0.05 mm, printing will not get you there consistently at volume, and you will pay for sorting. Molded parts hold tighter, and where they do not, we machine the critical features after molding.
Finally, look at the material. If the part needs UV stability, chemical resistance, or a flame rating, the molded grade list is longer and cheaper. If it needs a lattice or internal channel that a mold cannot pull, printing stays.
Bridge tooling and hybrid paths that buy time
The choice is not always print or production mold. There are middle steps. A soft tool or aluminum mold can run 1,000 to 10,000 parts and costs a fraction of hardened steel. That gets you into the market while the design settles. When the design freezes, the steel tool follows.
Machined parts are another bridge. For quantities in the hundreds to low thousands, CNC machining delivers molded-grade material with no tooling at all. Tolerance is ±0.005 mm and surface finish reaches Ra 0.8–1.6 μm. The unit cost is higher than molding but there is no upfront spend and no lead time for a mold.
A common pattern we see: print for the first 50 prototypes, machine 200 to 2,000 units for field trials, then mold once the design stops moving. Each step has a different cost curve and the handoffs are where projects get into trouble.
If you are unsure which step you are on, send the drawing and the annual volume. We will quote the print, the machining, and the molding path side by side so the comparison is on one page.
Seven steps to test whether you should move off additive
Run these in order. Stop as soon as one step gives you a clear answer.
- 1Pull the real annual volumeGet the number from the demand plan, not the PO history. Split it into design-stable and design-in-flux portions. A 30,000-unit annual volume that changes twice a year is not a molding candidate yet.
- 2Collect the fully loaded print costAdd machine time, material, support removal, bead blasting, and inspection into one number per part. If your current quote lists print time only, ask for the post-processing line items separately.
- 3Check the tolerance stackCompare the assembly requirement against what additive holds in production, roughly ±0.1 mm. Anything tighter than ±0.05 mm needs machining or molding plus a secondary operation.
- 4List the material requirementsWrite down UV exposure, chemical contact, operating temperature, and any flame or medical rating. If the list is long, molded grades cover more of it at lower cost per kilogram.
- 5Inspect for Z-axis riskAny part that holds pressure or carries sustained load should be flagged. Micro voids along the layer interface are the usual failure point. Pressure-test a sample before committing to volume.
- 6Get both quotes at the same volumeAsk for pricing at 1,000, 10,000, and 100,000 units for printing, machining, and molding. The crossover point is usually visible in the table without any further analysis.
- 7Plan the transition windowTooling takes weeks, and first-article approval adds more. Keep the print line running until the molded parts pass inspection, then retire it. Do not shut down the bridge before the new process is qualified.
Questions engineers ask before switching
At what quantity does injection molding for mass production become cheaper?
There is no universal number. It depends on part size, cavity count, material, and how much post-processing the printed version needs.
In practice, simple small parts often cross over between 5,000 and 20,000 units. Large parts with expensive tooling can sit above 50,000. The only reliable method is quoting both paths at your actual volume.
Can we keep printing some parts and mold others?
Yes, and many programs do exactly that. High-mix, low-volume spares stay on the print line while the high-runner goes to a mold.
The split usually follows revision cadence. Parts that change stay printed. Parts that are frozen move to tooling.
What tolerance can a molded part hold?
It depends on material shrink, wall uniformity, and gate placement. Well-designed parts hold tight tolerances in the flow direction and looser ones across it.
Where a molded feature needs to be tighter than the process allows, we machine it after molding. That keeps the tool simple and puts the precision where it matters.
How long does it take to move from printing to a molded part?
Tooling design, manufacture, and first-article approval typically run several weeks, and the schedule depends on part complexity and mold cavity count.
Keep your additive line running through that window. Production does not need to stop while the tool is being built.
Does surface finish change when we switch?
Yes. Printed parts carry layer lines unless they are sanded or vapor smoothed, and that is per part labor.
Molded parts take the surface of the cavity. A polished cavity gives a smooth part on every shot. Textures and graining are also added to the tool directly.
What if the design changes after the mold is cut?
Most changes are possible. Simple features can be welded and re-machined. Larger changes may need a new insert or a new cavity.
This is why we ask about revision cadence up front. A design that moves twice a year should stay on additive or bridge tooling until it settles.
Send the drawing and the annual volume
We will quote the print, machining, and molding paths side by side so you can see where the crossover sits for your part.
12-hour quote100% inspectionNo MOQ
More process notes from the shop floor
We publish setup notes, tooling trials and inspection data from the factory floor.
