SLS vs Injection Molding for Low-Volume Production: Where the Break-Even Actually Falls

Injection molding spends money before the first part exists. SLS spends it on every part. Which one costs less for a run of 50, 500 or 5,000 nylon parts comes down to three things: how far your quantity sits from the break-even point for that specific part, whether the design is frozen or still changing, and how many copies of the part fit in the printer at once.

Most comparisons stop at “3D printing is cheaper under about 1,000 units.” That figure is wrong often enough to be dangerous. Published cases on the same question range from a break-even of a few hundred units to one of 80,000, and the spread is not noise. It is what happens when you divide a tooling cost by a per-part saving that changes with every part. This article shows the arithmetic, runs it on three real SLS parts, and gives injection molding its due where it wins.

What each process is

Selective laser sintering (SLS) is a powder-bed process in which a laser fuses thin layers of polymer powder, usually a nylon, into a solid part, with the surrounding unfused powder holding the part up as it builds. No tooling is made, so the first part and the thousandth part cost roughly the same to produce, and the parts can be stacked through the whole build volume rather than laid out on a plate.

Injection molding is a process in which molten thermoplastic is forced under pressure into a machined steel or aluminum mold, cools, and is ejected, with the cycle repeating every few seconds. The mold is the expensive part, often four to five figures and several weeks to make, and once it exists each shot costs cents of material and seconds of machine time.

That asymmetry is the whole comparison. Everything below is a consequence of it.

Cost per part: where the break-even actually falls

Cost per part decides this comparison for most buyers, and it is the factor most articles get wrong by quoting a single break-even quantity. The break-even is not a property of SLS or of injection molding. It is a property of the part.

The arithmetic is one line:

Break-even quantity = tooling cost ÷ (SLS cost per part − molded piece price)

Below that quantity, SLS is cheaper in total. Above it, molding is cheaper, and the gap widens with every additional part. Worked once in plain numbers: a tool that costs $4,000, for a part that costs $1.00 on SLS and $0.15 molded, breaks even at $4,000 ÷ $0.85, which is about 4,700 parts.

Two things in that formula move a lot from part to part. The saving per part shrinks for small parts, because a small part uses little powder and many of them fit in one build, so the SLS price lands close to the molded price. The tooling cost grows for large or complex parts, because the mold needs more steel, more machining, and often moving sections to release undercuts.

The table runs the formula on three of the RMS220 reference parts Raise3D publishes. The SLS column is the published cost per finished part, covering material, labor, energy, machine depreciation and consumables at the stated number of parts per build. The molding columns are worked examples, not quotes. Published tooling prices for nylon parts of this size run from a few thousand dollars for a simple single-cavity tool to around $10,000 or more for a tool with cores or several moving parts, and the piece prices are typical of molded nylon at volume. Your toolmaker’s quote replaces them.

Part (material, size) SLS cost per part Parts per RMS220 build Example tool cost Example molded piece price Break-even quantity
Cable clip (PA11, 20 × 69 × 46 mm) $1.00 192 $4,000 $0.15 about 4,700
Ventilation duct (PA NEXT GB, 203 × 26 × 165 mm) $5.00 18 $9,000 $1.20 about 2,400
Intake manifold (PA12 GB, 180 × 116 × 160 mm) $20.70 4 $15,000 $4.00 about 900

The result runs against the common belief. The big, complex manifold tips to molding soonest, at around 900 units. The small clip stays cheaper on SLS the longest, past 4,000 units, because at a dollar a part there is almost nothing left for a mold to save. If your rule of thumb is “SLS for small batches of big parts,” the numbers say nearly the opposite: SLS holds its own longest on small parts in large counts, and on big parts only until the tool pays for itself.

Three caveats belong next to that table, and they cut both ways.

  • The break-even is cumulative over the tool’s life, not per order. A part you will need at 800 a year for four years should be compared at 3,200 units, which puts the duct on the molding side even though no single order comes near 2,400.
  • The manifold may not be moldable as one piece at any price. Internal channels and undercuts that SLS builds in a single part often need a multi-part tool, or a redesign into two molded halves plus assembly. The fair comparison is then one SLS part against a molded assembly, and the piece price above understates the molded side.
  • Small changes in the inputs move the answer a lot. Double the clip’s tool cost to $8,000 and break-even moves to about 9,400. Raise the molded clip price to $0.40 and it moves to about 6,700. Run a build half empty and the SLS cost per clip rises, pulling break-even down. Run the formula with your own quotes, not ours.

One more cost sits outside the table on each side. For SLS it is the system. The RMS220 printer with one build unit lists at $34,999, and a production setup adds a second build unit so one cools while the other prints, plus a cleaning station, a blasting station and dust extraction. For molding it is the mold’s own upkeep and storage, and the cost of a second tool when the design changes. Depreciation on the SLS system is already inside the published cost per part. The mold’s replacement cost is rarely inside anyone’s piece price.

Lead time: first part vs per part

Lead time matters because a cheaper part that arrives after the launch date is not cheaper. The two processes are fast at opposite ends of the run, and most comparisons only mention one end.

For the first part, SLS wins by weeks. On the RMS220, the ventilation duct above prints in 6 h 33 min, cools for 12 hours in the build unit, and is depowdered and blasted the next morning. Cutting a production mold takes four to eight weeks from approved design to first shot, and longer from an overseas toolmaker, before a single part exists. For a design still being validated, that gap is the difference between testing this week and testing next quarter.

For the thousandth part, molding wins by a similar margin. Once the tool is running, a press produces a part every few seconds. The RMS220’s rated output is 2.2 L/h at 20% packing and 3–5 kg of parts per day, so 1,000 ducts is weeks of continuous builds against an afternoon on a press. SLS output scales by adding build units and machines, not by running one faster.

The practical consequence is bridge production, and it is the one strategy every honest comparison lands on. Print the first few hundred units while the tool is cut, ship on schedule, then move the part to the mold when it arrives. The same CAD file feeds both, and a design error found in the printed batch is fixed before it is cut into steel. Automotive suppliers use this routinely when a customer changes a part too late for the tooling schedule.

Design changes and product variants

Design stability matters because a mold is a bet that the part will not change. Every revision after the tool is cut costs either a tool modification or a new tool, and the lead time clock restarts.

If the design is frozen, this factor is neutral. If it is not, it dominates. A product in its first year typically goes through fit changes, a wall thickened after a field failure, a boss moved to clear a new PCB. On SLS each of those is a file edit and the next build. On a mold each is a quote from the toolmaker, and some changes, such as adding material where the cavity already exists, cannot be made without welding or a new insert.

Variants are the same problem multiplied. A part sold in three sizes and two mounting patterns is six molds, or one mold with interchangeable inserts, and either way a tooling budget that grows with the catalog. On SLS the six variants nest in the same build, and a seventh costs nothing to add. Custom insoles, orthoses and operator-specific fixtures are the clear cases, but any product line with many variants and low volume per variant sits here.

The fair counterpoint: a locked design with high volume gets no benefit from this flexibility, and pays for it in per-part cost. Molding’s rigidity is only a cost when you need to move.

Geometry, nesting and part size

Part geometry matters because it sets the SLS cost per part through packing, and sets the mold cost through complexity, so the same feature can push the break-even in opposite directions at once.

SLS cost scales with the volume of powder a part consumes and the share of a build it occupies. The RMS220’s build volume is 220 × 220 × 350 mm, 17 L, and at the rated 20% packing density that is roughly 3.4 L of parts per build. A clip that fits 192 to a build shares the build’s fixed costs 192 ways. A manifold that fits four shares them four ways. That is the whole reason the clip costs $1.00 and the manifold $20.70 on the same machine with the same labor.

It also means the SLS cost per part is only valid at the stated packing. A production run of one duct in an otherwise empty build does not cost $5.00. Shops that get good SLS economics keep builds full, which is easier when the part mix is varied and the build unit swaps out for the next job while the last one cools.

Molding cost scales with the tool, and the tool scales with features. Undercuts need slides. Internal channels need cores or a split part. Thin walls, long flow paths and sink-prone thick sections need iteration on gates and cooling. None of this is impossible, and a good toolmaker solves it daily, but each item adds to the tool price and the lead time, and some geometries that SLS prints as one piece become two molded pieces and an assembly step.

Part size caps SLS in a way molding does not feel. The RMS220 builds to 350 mm in Z and 220 mm in X and Y. The headlight cover on the reference list, at 197 × 196 × 393 mm, prints one per build at $14.70 and uses most of the chamber. Anything larger is a larger machine or a split part. Molding handles that size as routine.

Materials and surface finish: where molding wins outright

Material and finish matter because they are the two places where no quantity calculation rescues SLS. If the part needs a polymer SLS cannot sinter, or a surface SLS cannot produce, the break-even is irrelevant.

Molding’s material list is the entire thermoplastics catalog: PP, PE, ABS, PC, POM, PBT, glass-fiber nylons, TPEs across the hardness range, flame-retardant and UV-stabilized grades, any color from a masterbatch, and clear parts. Overmolding and insert molding add a second material or a metal insert in the same shot. If a spec sheet names a specific resin grade, molding is usually the only way to get exactly that grade.

SLS on the RMS220 runs PA12, PA11, PA NEXT, glass-bead-filled PA12 GB and PA NEXT GB, and TPU at 86A and 90A, with an open-material license for selected third-party powders. That covers a large share of functional nylon and elastomer parts, and SLS nylon is roughly equally strong in every direction, which filament printing is not. It does not cover PP, PE, polycarbonate or anything transparent, and the native color is black or a dyeable grey. A part that must be glossy white, or match a Pantone, is molded.

Surface finish is the more common deal-breaker. A molded part leaves the tool with the finish machined into the cavity, from mirror polish to a specified texture, and every part matches. An SLS part comes out matte with a fine grain from the powder. Bead blasting, which the B520 does as a standard step, cleans and evens it, and PA NEXT is formulated to come off the blaster closer to a molded look, but it is a different surface and a trained eye tells them apart. Vapor smoothing closes the gap further at extra cost per part. For housings behind a panel, ducts, brackets, clips and fixtures, none of this matters. For a consumer-facing cosmetic surface it often decides the question on its own.

Tolerances and repeatability

Tolerance matters because a part that does not fit is scrap at any price, and the two processes get to a fitting part by different routes.

A validated mold is a tolerance machine. Once the first parts are measured and approved, the cavity, the pressure and the cooling repeat shot after shot for the life of the tool, and the commercial tolerance classes in standards such as DIN 16742 are routinely held on small features. The cost is paid up front, in the sampling and tool-correction rounds before production, and it is one reason tooling lead times run to weeks.

SLS accuracy is not a single number, and anyone who publishes one is simplifying. It depends on the material, the part’s geometry and wall thickness, where it sits in the build, how densely the build is packed and how the build cools. Service bureaus quote SLS tolerance as a percentage of the dimension with a floor of a few tenths of a millimeter, and tighter is achievable on specific features once the process is dialed in for that part. Flexible TPU moves more than PA11. The sensible approach, and the one Raise3D takes with customers, is to validate accuracy on representative sample parts before committing a production part to SLS, rather than trusting a datasheet figure.

In practice: snap fits, press fits and bearing bores work on SLS with a tolerance study and sometimes a reaming or drilling step. A molded part gets there without the study. Repeatability within an SLS build is good. Repeatability across builds depends on process discipline, powder refresh consistency and thermal control, which is where a machine with a stable process window and a controlled cooling cycle earns its keep.

The options in between

The choice is rarely binary, and a comparison that ignores the middle ground reads as a sales pitch. For runs in the hundreds to low thousands, at least four other routes compete with both in-house SLS and hardened-steel tooling.

  • Aluminum or soft tooling. A machined aluminum mold costs a fraction of a hardened steel one and is cut faster, at the price of tool life measured in hundreds to tens of thousands of shots rather than hundreds of thousands. For a frozen design at 1,000–10,000 units in a resin SLS cannot run, this is often the right answer, and this article will not pretend otherwise.
  • Printed or silicone molds. A mold printed in SLS nylon or resin and used for silicone or urethane pours costs a few hundred dollars and lasts a few dozen shots. That makes it the economical route for roughly 10 to 200 units of an elastomer part the TPU grades do not cover.
  • Urethane casting through a bureau. Silicone molds taken off a printed master, cast in urethane. Bureau pricing typically puts it ahead of both SLS and molding somewhere in the 30 to 250 unit band for mid-sized housings. The material is urethane, not nylon, which rules it out for some functional parts.
  • MJF or SLS through a service bureau. Outsourcing nylon parts avoids the capital cost entirely and is the right first step for most teams. The trade is price per part and lead time at the bureau’s mercy, plus a quote cycle for every design iteration. The point at which in-house SLS beats outsourcing is a cost-per-part question of its own: put the bureau’s price per part next to the in-house figure at the build utilization you will really achieve, and count the quote cycle time for every design change.

Where does in-house SLS sit among these? It wins when the part is nylon or TPU, the quantity is below the break-even for that part, the design or the range of variants is still changing, and the team prints often enough to keep builds full. When one of those is false, one of the options above is usually cheaper.

When SLS wins, when molding wins, when to run both

The decision is a checklist, not a verdict. Score your part against each row and the answer is usually clear by the third one.

Factor SLS is the better call when Injection molding is the better call when
Quantity Lifetime volume is below the break-even for that part (hundreds to a few thousand for most mid-sized nylon parts, higher for small ones) Lifetime volume is well above it, or annual demand is steady for years
Design stability Revisions are still expected, or the part exists in many variants The design is frozen and validated
Geometry Internal channels, consolidated assemblies, undercuts that would need slides Simple pull, uniform walls, large flat surfaces
Part size Fits a 220 × 220 × 350 mm chamber with room to nest Larger than a benchtop or mid-size SLS chamber
Material PA11, PA12, filled PA, TPU 86A–90A PP, PC, ABS, POM, specific resin grades, clear or colored parts
Surface Functional, hidden, or acceptable after blasting Cosmetic, glossy or textured to a spec
Lead time First parts needed in days Thousands of parts needed per week once running
Capital A system costing $35,000 and up will stay busy across many parts A single tool per part is the only investment wanted

Run both when the part will end up molded but is needed before the tool exists, when demand is unproven and tooling would be a gamble, or when a molded product carries a few low-volume variants that are not worth their own tools.

Frequently asked questions

At what quantity does injection molding become cheaper than SLS?

There is no single number. Break-even equals tooling cost divided by the gap between SLS cost per part and molded piece price, so it ranges from under 1,000 units for a large complex part to several thousand for a small clip. Run the formula with your own quotes. Published cases span a few hundred units to around 80,000.

Is SLS nylon as strong as injection-molded nylon?

Close, for PA12 and PA11. SLS parts are roughly equally strong in every direction and sintered to high density, so tensile and bending strength sit near molded figures for the same polymer, with stretch before breaking usually lower. Fatigue and impact performance should be tested for the specific application rather than assumed from the datasheet.

Can SLS parts replace molded parts in end-use products?

Yes, and they routinely do for brackets, housings, ducts, clips, fixtures and custom-fit components in nylon and TPU. The limits are cosmetic surface, materials outside the nylon and TPU families, part size beyond the build chamber, and per-part cost once volumes climb past break-even.

How does SLS compare to FDM for low-volume production?

SLS needs no support structures, stacks parts through the whole build volume, and produces nylon that is equally strong in every direction, so it delivers far more finished parts per machine per day than FDM and parts that behave more like molded ones. FDM is cheaper to buy and run for one-offs and large single parts.

What does an in-house SLS setup cost compared to a mold?

The RMS220 printer with one build unit lists at $34,999, and a production setup adds a second build unit, cleaning and blasting stations and dust extraction. A single mold runs from a few thousand dollars for a simple cavity to tens of thousands for a cored or multi-cavity tool. The difference is that the SLS system serves every part you make, and the mold serves one.

Is SLS worth it if I only need a few hundred parts a year?

For a few hundred parts of one design, outsourcing SLS or using cast urethane is usually cheaper than owning a system. In-house SLS pays off when there are many different parts, frequent revisions, or customization, so the machine stays busy. The deciding figure is how many hours a month the printer would actually run.

Can I start on SLS and move to molding later?

Yes, and it is the most common path for new products. The same CAD file feeds both processes, with design-for-molding adjustments such as draft angles and uniform walls added when the tool is designed. Printing the first few hundred units validates the design before it is cut into steel.

Conclusion

For low-volume and small-batch nylon parts, SLS is usually the lower-cost route, but “usually” hides a break-even that moves by an order of magnitude with the part. A small clip at a dollar a piece stays on SLS past 4,000 units. A large manifold at twenty dollars a piece tips to molding before 1,000, if it can be molded in one piece at all. Divide your tooling quote by the per-part gap and you have your own answer in thirty seconds.

Injection molding keeps every advantage it has always had: cents per part at scale, any thermoplastic in any color, a cosmetic finish from the tool, and tolerances held for the life of the mold. What has changed is the price of admission to SLS. A system like the RMS220 brings production nylon in-house for less than the cost of a handful of molds, which turns the decision for each new part from “can we afford a tool” into “is this part past its break-even yet.” That is a better question, and it has a different answer for every part on your list.

If you want to run the formula on your own parts, the inputs you need are a tooling quote, a molded piece price, and an SLS cost per part at a realistic packing density. Raise3D’s application engineers will produce the last of those from your CAD files, along with sample parts to check fit and finish, before you commit to either route.

More Content