Injection Molding vs 3D Printing: Cost & Quality Crossover

Injection molding and 3D printing answer the same need, a plastic part, with opposite economics. Molding buys a tool once and then pays a few seconds of press time and a little material per part, so its cost per part falls steeply as the run grows. Printing pays no tool but roughly the same for every part, so its cost per part stays nearly flat. Where those curves cross is the decision, and the commonly cited crossover is roughly 250 to 2,000 parts: below it printing usually wins on cost, above it molding usually wins.

That band is an anchor and a moving target at once: part size, tooling class, and which printing process you quote against can slide it by an order of magnitude. This page works the break-even arithmetic, the tolerance and strength data behind the quality claims, and the tooling ladder between no tooling and production steel. The short version: print while the design moves, bridge-tool in the production material once the design freezes, and cut production steel at volume in the thousands.

Two processes, two cost structures

The two processes form parts by opposite routes, and the mechanics explain the economics. Injection molding melts thermoplastic pellets and injects the melt into a machined steel or aluminum cavity under high pressure, where it cools and ejects as a finished part. Draft angles, wall thickness, gate location, and ejection strategy all live in that cavity, decided weeks before the first part exists. Change the part and you change metal.

3D printing builds the same part additively, layer by layer, with no cavity anywhere. FDM extrudes melted filament along a toolpath, SLA cures liquid photopolymer with light, and SLS and MJF fuse nylon powder across a build bed. A printed part exists as data until it builds, so a design change costs a reprint, and internal channels, lattices, and consolidated assemblies no mold could form are ordinary print work. The trade: material breadth, as-built finish, and layer weakness.

The comparison table later in this page follows from those mechanics. The injection molding and 3D printing hubs cover each process in depth; this page stays on the three rows that decide most purchases: cost, accuracy, and tooling life.

Where the cost crossover sits

A molding quote carries two parts, the tool as a one-time cost and the parts as material plus seconds of press time each; a printing quote carries almost only parts, with no tool to amortize. That fixed-versus-per-part split is the anatomy of every manufacturing quote. The break-even quantity follows from one division, commonly reasoned as: divide the tooling cost by the per-part saving that molding delivers over printing. Every part below that quantity is cheaper printed, every part above it cheaper molded. An expensive tool pushes the crossover out; a large per-part saving pulls it in.

For a mid-size part with a conventional single-cavity tool, the commonly cited crossover is roughly 250 to 2,000 parts, an anchor, not the answer. A simple part in a commodity resin, tooled cheaply in aluminum, can cross in the low hundreds; one worked case had a 4,800-piece nylon order whose tool paid for itself at around 130 sets, under 5 percent of the run. Complex geometry, large parts, or multi-cavity hot-runner steel tooling push into the several thousands, because tool cost grows faster than the per-part saving.

Between roughly 500 and 1,500 parts sits a gray zone where unit cost alone stops deciding; tooling class, cash available, and the value of lead time decide instead, which is why this page treats the crossover as a ladder of tooling bands rather than a single number.

Which way your part pushes the break-even matters more than the band itself. Time-to-market pressure, a live probability of design change, undercuts and internal channels, and materials that will not mold push the break-even up and favor printing longer. A frozen design, commodity resins such as ABS, PP, PC, and PA, continuous production, and unit-cost targets only high volume can reach push it down and favor tooling sooner.

Which printing process you measure against matters too. FDM is the cheapest printed route for simple geometry, while SLS and MJF nest many parts into one powder bed, so their per-part cost falls with batch size. Their break-even against molding usually sits in the tens to low hundreds of parts, and the printed cost drivers behind it are covered on the 3D printing quote page.

Lead time crosses over too. Printing delivers first parts in days; molding delivers nothing until the tool exists, commonly weeks, then produces parts in seconds each, so per-part lead time falls steeply for molding as volume rises, a crossover the manufacturing lead time benchmarks page tracks in relative terms.

Accuracy: a claim worth correcting

Plenty of pages on this topic claim 3D printing is the more accurate process, and for the printing workhorses that claim does not survive contact with the tolerance bands. A molded part typically holds about ±0.1 to 0.3mm, tighter within one mold half and looser across the parting line, with critical features held by tighter mold tolerancing. ISO 20457 governs tolerances and acceptance conditions for molded parts and requires measurement in a stabilized state, because parts keep shrinking after ejection and read larger at the press than two days later.

Printed tolerances are process-bound, so one number for 3D printing misleads. FDM runs about ±0.1 to 0.5mm depending on material, ABS at the tight end and flexible TPU at the loose end. SLA holds about ±0.05 to 0.15mm, the only print process that matches and sometimes beats a mold, and only on small parts. MJF runs about ±0.2 to 0.3mm, and SLS about ±0.3mm for features under 50mm and 0.5mm above that.

Ranked honestly, SLA can out-tolerance molding on small, detailed parts, while FDM, SLS, and MJF, the three that print functional parts in bulk, are equal to or looser than a good mold. The accurate claim is narrower: 3D printing wins on geometric freedom, not dimensional tolerance. Per-process detail sits on the FDM vs SLA vs SLS page, and features tighter than any mold holds belong on a machine tool.

Surface finish follows the same shape. A mold produces finish in the cavity, from a diamond-polished SPI A gloss to a dry-blasted matte or an etched texture, so every part in a run is identically finished with no per-part labor. A printed part carries its process signature as-built: SLA around Ra 0.5 to 2 micrometers, MJF around Ra 4 to 9, FDM Ra 4 to 12 with visible layer lines, SLS Ra 5 to 12. Cosmetic or sealing surfaces on printed parts need per-part finishing, labor that scales with quantity; the surface finish Ra reference table holds the numbers.

Strength, isotropy, and material breadth

Molding forms the part as one coherent solid under pressure, so a molded part is effectively isotropic, the same properties in every direction. FDM prints are not: the bond between layers is weaker than the material within a track, leaving a part about 20 to 30 percent weaker across the Z, or build, direction than in-plane, a gap you design around by orienting loads along the layers. SLS and MJF nylon parts are near-isotropic, the closest printed equivalents to molded strength, which is why load-bearing printed work routes to them.

Material breadth favors molding heavily. A mold runs essentially the whole commodity and engineering thermoplastic catalog, PE, PP, ABS, PC, PA, POM, TPU, and glass-filled grades, and certified grades with property sheets are overwhelmingly molding resins. Printing is process-bound: FDM covers the filament families, SLS and MJF are essentially nylon-family, and SLA is photopolymer only. If a specified resin with a data sheet matters, check the engineering plastics page before assuming a printed substitute.

The tooling ladder: three bands, three shot lives

Between no tooling and production tooling sits a middle band most comparison pages skip, and it is where the crossover gets decided. Think of the journey as three rungs, each with its own tool life in shots; a shot is one molding cycle.

Band one is no tooling: print the parts. It fits roughly one to a few hundred parts, or any volume where the design is still moving; the rapid prototyping page covers that work. First parts arrive in days, geometry stays free, and the constraints are the ones above. Printing the early parts also de-risks the tool: parts and assemblies see real use before any steel is cut, a point the low-volume manufacturing page develops.

Band two is bridge tooling: a real metal mold, usually aluminum or P20 pre-hardened steel, built to production intent, targeting SPI Class 104 in aluminum and Class 103 where P20 is used, and delivering roughly 10,000 to 100,000 shots. Aluminum cuts faster and dissipates heat better than steel, so both the tool build and the molding cycle run faster, which makes it the default bridge material. The decisive advantage is continuity: bridge parts come out in the production material with production-grade finish and repeatability, so validation carries over unchanged.

Band three is production tooling: hardened steel, multi-cavity and hot-runner capable, running into the millions of shots, with the lowest per-part cost on the ladder and the longest lead time and highest change cost. A design change at this rung can mean re-cutting or scrapping a cavity, which is why nobody should climb it while the design still moves.

Shot life separates the rungs, and it is one of the three inputs, with part size and material, that set where your crossover lands. As a common ordering: printed plastic molds last tens to about 100 shots, aluminum and P20 bridge tools roughly 10,000 to 100,000, hardened production steel millions.

The printed mold trap

Printed injection molds, photopolymer cavities from SLA-type systems, get marketed as rapid tooling. A printed plastic mold typically becomes ineffective within about 100 shots on soft, hot resins like PE and PS, and yields only a handful of parts from glass-filled or high-temperature materials. Part volume caps around 10 cubic inches (164 cm³), draft needs 5 degrees or more against the 1 to 2 a metal tool asks, gates must be enlarged, pressures lowered, and cycles stretch because plastic cannot dissipate heat. The cavity still needs holes drilled, threads tapped, and layer lines sanded. Treat printed molds as a prototyping aid, not a bridge strategy.

Where additive manufacturing genuinely belongs in tooling is metal inserts. In one peer-reviewed case study, a laser powder bed fusion insert with 32 cavities proved as capable as a conventionally machined 16-cavity insert over an assumed 10-million-shot tool life, with comparable cycle times, on conformal cooling channels only AM can form. Production tooling made additively is a different thing entirely from a printed plastic mold.

When to choose which

Three questions decide most real cases. Is the design frozen? Is the quantity stable at or above the low hundreds? Does the part need production material properties, a cosmetic finish as-made, or part-to-part repeatability at that volume?

Choose 3D printing while any answer is no. That covers prototypes and fit checks, jigs and fixtures, any part whose geometry is still under revision, volumes below the crossover, and internal channels or consolidated assemblies a mold cannot form without side actions.

Choose injection molding when the design is frozen, the volume is stable in the hundreds to thousands or beyond, and the part needs isotropic strength in a specified resin, a finished surface straight out of the cavity, or certification only molding resins carry. Designing for the mold differs from designing for the print: draft, uniform walls, and gate and ejection strategy are mold constraints printable geometry is free to ignore, and the design for manufacturing page covers the difference.

Bridge production: running both at once

Bridge production is running a tooling-free or soft-tooled process to ship parts while the high-volume tool is being built, the honest answer to most gray-zone quantities. The bridge parts cost more each, but they buy time to market and de-risk the tooling, because the parts and the assembly see real use before the steel is cut.

Two paths work in practice. Print the bridge parts, usually SLS or MJF for functional work since both nest parts into one bed, while the production tool is cut. Or cut an aluminum bridge tool in the production resin and run it until the hard tool is ready, which keeps material, finish, and validation continuous across the transition.

The sequencing rule is short. If the design is still moving, print. If the design is frozen and the volume sits in the gray zone, bridge-tool it in the production material. If the volume is comfortably in the thousands and the design is frozen, tool for production and run a bridge in parallel to protect the launch date.

When neither process fits

Some parts leave both processes behind. Features tighter than about ±0.05mm belong on a machine tool. A metal housing at volume belongs to die casting, the tooling-amortization mirror of molding for metals. One simple part is often cheapest machined from plate, and very large or thick geometry can exceed both printing build volumes and molding economics, where machining or fabrication from stock wins. The manufacturing process comparison matrix page rounds out the whole-process view.