A founder with a finished design and a first order for 300 units faces an uncomfortable quote: $28,000 for an injection mold, twelve weeks lead time, and a per-part cost of $1.40. Spread over 300 units the tooling alone is $93 a part. Spread over 50,000 units it is 56 cents. The mold is the right answer eventually and the wrong answer now, and the art of low-volume manufacturing is bridging that gap without spending capital you do not have on volume you cannot yet sell.
A mature set of no-tooling and low-tooling processes exists, each with a different sweet spot. They are not interchangeable, and choosing badly is expensive in a quieter way than tooling too early.
The Options, Honestly Compared
| Process | Tooling cost | Typical per-part | Sweet spot | Main limitation |
|---|---|---|---|---|
| SLS / MJF printing | None | $8–$120 | 1–500 units | Nylon-family materials only; surface texture |
| Urethane casting | $400–$1,800 per silicone mold | $40–$300 | 25–500 units | Mold wears out after 20–40 pulls |
| CNC machining | Fixturing only | $25–$400 | 1–300 units | Cost barely falls with volume |
| Sheet metal | None for laser and brake | $15–$200 | 1–5,000 units | Geometry limited to bends and cutouts |
| Thermoforming | $500–$4,000 | $5–$60 | 50–5,000 units | Open shapes, one good side, uneven wall |
| Aluminum bridge tooling | $3,000–$15,000 | $1–$8 | 500–30,000 units | Limited tool life; fewer resin options |
| Steel production mold | $15,000–$120,000 | $0.30–$5 | 10,000+ units | Capital and lead time |
Urethane casting
Also called vacuum casting or silicone molding. A master pattern, usually printed or machined, is suspended in liquid silicone; once cured the silicone is cut open and used as a mold for two-part polyurethane resin poured under vacuum. Parts look and feel remarkably close to molded plastic, take color in the resin, and can imitate ABS, polypropylene, clear acrylic, or rubber-like durometers.
The economics are ruled by mold life. A silicone mold yields roughly 20 to 40 parts before detail degrades, so 200 parts means six to ten molds, which is why per-part cost stays nearly flat with volume. Cycle time is hours per part per mold, so schedule accordingly. Undercuts are far more forgiving than in injection molding because the flexible mold peels away.
3D printing for end-use parts
Powder-bed nylon processes are the workhorse here. SLS and HP's MJF produce isotropic, functional parts with no support structures, good living-hinge and snap-fit behavior in nylon 12, and consistency good enough to ship. Cost is driven by the volume of powder your part occupies and how densely parts nest in the build, which means design changes that shrink or hollow the part translate directly into money. The category is broad enough to deserve its own treatment; see when printing beats tooling for production parts for the material and finish tradeoffs.
CNC machining
Machining wins when you need real engineering material properties, tight tolerances, or a structural part. The cost curve is close to flat: a run of 200 costs nearly 200 times a run of one, minus setup amortization. Design choices move the number far more than quantity does, which is why what actually drives the price of a machined part is worth reading before you send the quote package.
Sheet metal
The most underused low-volume process. Laser cutting and press-brake forming need no dedicated tooling at all, hardware can be pressed in, and the finish options are mature. If your enclosure can be expressed as flat panels and bends rather than molded curves, sheet metal will typically beat every plastic option below a few thousand units. It also scales upward gracefully. The design rules that keep it cheap are in bend radii, reliefs, and sheet metal DFM.
Thermoforming
For large, shallow, open shapes such as covers, trays, and housings for equipment, vacuum forming beats injection molding badly at low volume. Tooling is a single-sided pattern rather than a two-part steel mold, and it can be machined from aluminum or even a filled composite. The constraints are real, though: wall thickness thins where the sheet stretches, and only one surface is controlled. See vacuum forming and thermoforming for low-volume parts.
Bridge and soft tooling
This is the middle ground and it is where most hardware companies eventually land. An aluminum mold cuts faster than steel, costs a fraction as much, and can run tens of thousands of shots in a forgiving resin. Master Unit Die inserts go further: you buy only the small cavity insert and rent the standard frame it drops into, cutting cost again. Single-cavity, hand-loaded, unautomated versions of production tools are the usual first step. The tradeoffs are covered in choosing between aluminum and steel molds.
The Crossover Math
The decision is arithmetic, not opinion. Compute the breakeven quantity between any two options:
Breakeven = (Tooling A − Tooling B) ÷ (Per-part B − Per-part A)
Take a housing quoted at $80 per part in urethane casting with $1,200 of silicone molds, against $12,000 for an aluminum bridge tool running at $3.50 per part. The difference in tooling is $10,800; the difference in per-part cost is $76.50. Breakeven is about 141 units. Below that, cast. Above it, cut the tool.
Three corrections keep this honest:
- Include the cost of being wrong. Tooling is spent before you know the design is final, so the expected cost of a tool includes the probability of modifying or rebuilding it.
- Include lead time. Cast parts in three weeks versus molded parts in fourteen may be worth more than the per-part difference if it means hitting a season or a customer commitment.
- Include finishing labor. Low-volume processes usually mean higher scrap and more hand work, which quotes sometimes omit. Read them using what the line items in a manufacturing quote actually mean.
A Sensible Sequence
- Units 1–25: print or machine everything. Iterate the design freely; there is no sunk tooling to protect.
- Units 25–300: urethane casting for cosmetic plastic parts, sheet metal for structure, machining for anything load-bearing. This is the pilot run where you learn assembly and find the design faults that only appear in quantity.
- Units 300–5,000: bridge tooling for the highest-volume plastic parts, keeping low-count parts in the low-volume processes. Freeze the geometry that goes into a tool first and keep the uncertain parts flexible.
- Beyond that: production steel, multi-cavity where the volume supports it, and a real conversation about automation.
The general framework for matching process to quantity is set out in how to choose a manufacturing process by volume, and the full picture of what a mold costs and returns is in injection molding costs and cheaper alternatives.
One discipline pays across all of this: parts drawn to be moldable from the beginning move between casting, bridge tooling, and steel tooling with minor changes. Parts drawn only for printing usually have to be redesigned, and that cost arrives at exactly the wrong moment.
Projects House plans low-volume production paths that scale, sourcing the right process for each part and keeping the design ready for the tool you will eventually cut. Tell us your volume and timeline through our contact form.