You have a validated prototype and a purchase order for the first production batch — 150 units next quarter, maybe 1,200 by year end if the product sells. Now comes the decision that quietly decides your margin for the next year: which manufacturing route do you commit to? Pick a process sized for volume you do not have yet, and you will pay for capacity that sits idle. Pick one that cannot scale, and you will rebuild everything halfway through.
There is no single right answer, but there is a repeatable way to think about it: match the process to the quantity band, and pick routes that degrade gracefully when volumes shift. Here is how the options actually compare across the 10-to-10,000 range.
The 10–150 Zone: CNC Machining Still Wins More Often Than Expected
Machining has a reputation for being expensive per part, which is true at 10,000 parts and misleading at 100. The reason is that CNC machining carries almost no fixed cost — no tooling, no mold, no minimum order. Your cost curve is nearly flat, driven mostly by machine time and setup. At 10 units a machined aluminum housing might cost 3–5 times the eventual die cast price; at 150 units the gap narrows enough that total program cost (including tooling amortization) often still favors machining.
Machining also buys you something buyers chronically undervalue: iteration. If the first 50 units reveal that a boss is wrong or a seal seat needs to be 0.2 mm wider, the fix is a program edit, not a mold modification. In the early production phase, design changes are the norm, not the exception — and every change that hits a tool costs real money and two to three weeks.
- Best fit: under roughly 150–200 parts, or when tolerances tighter than ±0.1 mm are concentrated in the part.
- Watch out: very thin walls (under 1 mm in aluminum) and deep pockets drive machining time nonlinearly.
- Batch trick: fixturing for 10 parts at once can cut per-part machine time substantially — ask about multi-station fixtures before assuming the quote is fixed.
The 50–500 Zone: Vacuum Casting and Soft Tooling Fill the Gap
Between "too many to machine economically" and "not enough to justify an injection mold" sits the awkward middle — and this is where vacuum casting (RIM) earns its keep. A silicone mold costs a fraction of a steel tool and delivers polyurethane parts in a few days, with surface quality that can match injection molding and even overmolded soft-touch layers. The trade-offs are real: PU materials approximate but do not equal ABS, PC, or glass-filled grades, dimensional tolerance is typically around ±0.2 mm on small features, and heat resistance usually tops out well below engineering thermoplastics.
For housings, cosmetic covers, and user-facing parts in the 50–300 range, that compromise is often acceptable — especially when the alternative is waiting six weeks for steel. For anything structural, high-temperature, or creep-sensitive, prototype-grade PU will mislead you, and you should budget for a real material path sooner. Where the volume outlook is firm, soft tooling with aluminum injection molds is the natural next rung: prototype molds in aluminum can run tens of thousands of shots in many materials, at cycle times and part properties close to production tooling.
The 500–10,000 Zone: Rapid Tooling and Die Casting Enter the Picture
Around a few hundred to a thousand units, the economics flip. Now a permanent tool's fixed cost amortizes to something small per part, and the per-part advantage of molded or cast processes — 30-second injection cycles, 60–120 second die casting cycles — becomes decisive. A plastic part that cost $8 machined and $4.50 vacuum cast may drop below $1 once an injection mold is amortized over several thousand shots.
For aluminum parts in this band, die casting becomes the default if the geometry cooperates: uniform walls, draft angles of 1–3 degrees, generous fillets, and no features that demand heavy secondary machining. The tooling lead time of 4–8 weeks is the price of admission, so the successful pattern is to overlap tool build with CNC-machined or vacuum-cast bridge parts covering initial demand.
Two numbers worth negotiating early, because they shape the whole program:
- Amortization structure. Tooling paid upfront versus amortized into piece price over a committed quantity changes your cash flow more than the piece price itself. A $15,000 tool amortized over 5,000 shots adds $3 per part; over 30,000 shots it is $0.50.
- Tool ownership and transfer. Insist the tool is yours after payment and can move to another molder. It is your insurance against single-supplier lock-in.
Designing Once, Scaling Anywhere
The biggest cost lever in low-volume production is not process selection — it is designing the part so the process can change. A few habits make the 10-to-10,000 journey survivable:
- Keep wall thickness uniform and moderate (2–3 mm plastic, 2–4 mm aluminum). It suits machining, molding, and casting alike, and it is the single most common fix for sink, warp, and porosity defects downstream.
- Design the CAD for a machinable version even if you intend to mold: avoid features only a specific process can produce, so the bridge batch and the final process can share geometry.
- Put critical tolerances on features that survive every process. Deep-pocket, thin-rib, and long-uniform-wall features tie your hands later; datum-friendly geometry keeps every option open.
- Plan the material migration path early. If final intent is glass-filled PA66, test the concept in that material family as early as you can afford — swapping material late reopens every mechanical question.
A Simple Decision Sequence
When the volume forecast is uncertain — and it always is — run this sequence: prototype and first articles in machining; bridge volumes of 50–300 in vacuum casting or soft tooling if cosmetic quality matters; commit to rapid or production tooling only when cumulative demand is credible above roughly 500–1,000 units; and keep the bridge process qualified until the new tool has survived its first quality escape-free month. Staggered like this, the expensive commitments happen only when the data supports them, and no single decision can sink the program.
Low-volume production is less about picking the perfect process and more about sequencing imperfect ones well. If you want a second opinion on where your quantities currently fall, send the drawing and forecast — our prototype services span the whole range, and a 20-minute review usually surfaces at least one step in the cost curve worth removing.
