I've been 3D printing as a hobby for a couple of years now, and have always heard how 3D printing is great for prototyping, but once you get beyond 10 or 20 pieces it's just not cost competitive with other technologies. I just accepted this to be truth, and have even told some of the customers I've gotten through 3DHubs and other marketplaces the same thing.

Now I'm creating an item for resale and am 3D printing the case. As I'll need 40 - 50 cases at a time I thought I would shop around for other options. But I just received some quotes back on injection molding. The mold price (for half the case) was \$15k, plus \$10 per unit at a quantity of 50 and \$4 per unit at a quantity of 500. Even in the case of the latter, the cost of tooling plus production is \$34/piece. 3D printing the same part myself costs $7 in filament, and paying someone else to 3D print it cost \$28.

Keeping in mind that 3D printing allows me to make changes to the design on the fly and print the exact number I need without having to worry about volume discounts, I now wonder if I've just been repeating a common misconception. Is 3D printing really noncompetitive for medium (10-500 pieces) production runs? If so, what makes it noncompetitive, given the extremely high up-front cost of IM?

  • 2
    $\begingroup$ Equating "filament cost" to total cost is a shortcut to going out of business. Printer & maintenance costs money, power to feed the printer costs money, space the printer takes up in house or shop costs money, other work the printer could be doing is lost money while printer is occupied... It's a trick going from a hobby mindset where that's all "free" to actually not bleeding money in a business setting; a trick many would-be businesses don't manage well. $\endgroup$
    – Ecnerwal
    Apr 22, 2017 at 3:02
  • $\begingroup$ @Ecnerwal Yes, that's true, but that's why I included the cost that I was quoted for someone else to 3D print it, which is 4x the filament cost. That is still less than the injection molding cost per piece, which is why I thought it odd. $\endgroup$
    – Nicholas
    Apr 23, 2017 at 15:48
  • $\begingroup$ In your particular example, unless you can find a place to do the molds for less, the break-even is a bit further out - once it comes, it comes on pretty strong, though. Without bothering to work it to the finest detail, somewhere before 650 is break-even (assuming per piece cost remains at $4) - and then the more you need the less it costs. Whether the other 3D printing operation has got their pricing model correct is unknown (I suppose you could order 100 parts and see if they suddenly want to renegotiate after printing a few dozen, but you'd likely prefer to run your own printer instead.) $\endgroup$
    – Ecnerwal
    Apr 23, 2017 at 18:38
  • $\begingroup$ See also 3D printed mold techniques for long and short term usage $\endgroup$
    – Greenonline
    Jun 25, 2017 at 5:52

4 Answers 4



3D printing is great for low quantities of items, but terrible for large quantities of items. This is a continuous scale and as your quantity grows, the drawbacks of 3D printing become more pronounced. In your situation, you seem to be dancing in the nebulous zone where there isn't really a good answer as to which production method is better. It's more of a question of price vs speed.

There are several factors that limit additive manufacturing technologies (in this answer I'll just talk about FDM since you're asking about a plastic item) from being practical in large scale manufacturing.


Compared to injection molding, 3D printing takes a very long time. An injection mold can make 72 plastic caps in 3 seconds. While I don't have a source, I would expect a similar item to take at least 5 minutes to be printed. In order to produce an average of 24 caps/second, you would need 7,200 3d printers working without down time. That doesn't take into consideration ...


Every 3d printed object needs to be removed from the print bed by hand. That requires a person to be there ready to remove the item. It's possible to have automated print removal, but for now I've yet to see it as a standard option on any machine. An injection model machine can just pop the parts out into a box. Even with the capability to remove the parts extra labor is still required for 3d printing because ...


Injection molded parts are as smooth as the mold they were made in. While failures still exist even for molded items, their relatively cheap production cost and very quick production time ensure that an occasional failure won't be problematic. 3D printing on the other hand is still prone to variations in the final product. For a few items, it's sensible to hand finish the parts with filing, sanding, or other reductive manufacturing techniques, but for many items this quickly become burdensome. Also, given the long production time compared to injection molding, even 1 discarded product will result in a long time required to produce a replacement.

  • $\begingroup$ Belt printers inherently have automated print removal. :-) $\endgroup$ Aug 5, 2020 at 22:58

Short Version

3d Printing is generally the most expensive method for building things mainly because it takes a long time to do. Compared to other methods it's by far the slowest. The few machines that can do parts quickly typically have very small build volumes so they're limited to only a few parts. Compared with injection molded parts you can make a dozen parts in a few seconds 3D Printing is closer to a few hours for one or two parts. Unless there is a change in the process or the technology or philosophy of the process 3D printing will always be delegated to small runs.

Unspecific Answer

You implied something in your post that you didn't explicitly ask for which I might be able to help. 3D printing is small scale, Injection molding is a large scale, but you've glossed over the medium scale production methods that might be worth considering given your project. If you're wanting to make small runs (20-500) units, your time could be well spent looking into cast-urethane parts. A mold will run around 200-500\$ depending on manufacturer, complexity, and size of the mold each mold is typically good for about 20 parts and then materials are roughly 20-30\$ for smaller sized parts. One of the more attractive aspects of cast urethane is that the material properties are very comparable to an injection molded part and the surface finish looks fantastic

In my experience, my workflow for a new part design is 3D print a few models to get from revision 0 to something you're confident should work, cast urethane a few batches to get functional field testing completed, and possibly even sell while the injection molding side gets set up. Injection molding takes 5-6 weeks about as a minimum for the first article, 3D printing is a few hours, and cast urethane is 2-3 weeks tops, less if you pay a bit more.

edit Forgot to mention that by using cast urethane as an intermediate you still have the flexibility to change your design without resulting in huge mold redesign costs. Because the mold is only good for about 20 parts modifying you can modify your part and update without incurring huge re-molding costs.

  • $\begingroup$ I like this answer as it directly addresses the fact that even though everyone says other methods should be cheaper, my experience directly counters that. I'm curious about the cast urethane though. You said they're only good for 20 parts; does it then break down and a new mold is needed? If so, that's $500 / 20 = $25 per part just for the mold. I'll look further into this and try to get some quotes; thanks. $\endgroup$
    – Nicholas
    Mar 27, 2017 at 13:58
  • $\begingroup$ Definitely get a quote pricing could vary a great deal depending on who you're talking to, the process they use and of course part complexity. It's been my experience that a mold is good for about 20 parts though if it's a simple enough part I think they could last for more runs. Whatever the number is, the mold is effectively useless at that point and it's discarded so a new mold would have to be made then. The molds are typically made from silicone so they're much easier to damage and the constant pulling of the part tears them up as I understand it. $\endgroup$
    – Diesel
    Mar 27, 2017 at 14:49
  • 2
    $\begingroup$ The moulds may only last 20 parts, but the master used to cast the mould should be good for casting more than one mould. $\endgroup$ Jun 27, 2017 at 11:34

Some small start ups (mainly 3d printing companies) do in fact use 3D printed parts for the production printers they sell. So, at least in some instances, 3D printed parts may be satisfactory for production products.

That being said, other manufacturing methods (such as injection molding) currently have several advantages over 3D printing.

Injection Molding

Pros: It can produce parts of different sizes ranging from small, to rather large (like a plastic chair). It provides superior inter-material bonding, takes only a few seconds to produce a single part, and can huge quantities of parts (possibly millions) in a single week. Typically injection molding is often entirely automated, with few moving parts. Small details can be repeated well and parts can be made with many different polymers.

Cons: Some geometries (like a hollow sphere) which can easily be produced with 3D printing, are impossible to produce with injection molds. Start up costs very be high. A typical machine can cost hundreds of thousands to millions of dollars. Mold cavities are costly to produce and require injection molding artifacts to be designed into the part (like ejector pin notches). Also, the mold produces edge lines on the part where the two cavities meet. Defect trouble shooting can be very complex and the ability to produce satisfactory parts can depend on heat flow (which can be affected by things like the weather). This can require climate control to eliminate to reduce waste in scrap parts due to warping and other malfunctions. Sometimes advanced statistics are required to measure dimensions and detect developing problems too complex to deductively trouble shoot because they stem from a fluid thermodynamic issues.

Once 3D printing has time to develop further (as injection molding has), we may see it used more and more for production parts.

  • $\begingroup$ Also for some injection molds, they can be very large in general (like 10 wide x 15 long x 12 feet high) and in comparison to the part they produce. $\endgroup$ Mar 24, 2017 at 15:51
  • $\begingroup$ Also I have seen injection mold producing around 70 million parts a year (run 24/7 days a week until break down for several weeks) with around 2-3% yield in actuality though they reported around 1% cause they lie. So that is about 1 million scrap parts. I think this is a good estimate on the yield injection molds can obtain here in the states. China and west/northern Europe probably below 1%. $\endgroup$ Mar 24, 2017 at 15:57
  • $\begingroup$ I meaned scrap, so yield -100 $\endgroup$ Jun 26, 2017 at 11:59

Break even points have always been dependent on the part. If the goal is to make a rod or filament-like piece of your extrusion nozzle diameter, your extruder may be pretty competitive. Additive manufacture has come a long way already so the break even points may have shifted. Interestingly a great use for additive is to produce tooling.

You could for example use the 3D printed part (with non-stick coating) to make a 600 deg F silicone mold from high temp RTV. Presuming you have the heating equipment, that mold can then be used to produce tens of the part at slightly lower quality. As silicone is flexible, it isnt good for high speed/pressure of injection molding.

If you are willing to accept lower quality (3d printing with filament is pretty low so that's likely a yes) and your injection molder is willing to go for DMLS molds with minor smoothing/drilling post-operations, your tooling costs may be significantly lower than your original estimate. Chances are this isn't happening unless the injection molder is also an additive enthusiast, and often will have to be one and the same as the person who designed at least some of the 3d printed part.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .