Quantity decides most of it: 1–50 pieces are printed or machined, 1,000+ are molded, and the 50–1,000 band in between — the one most comparisons skip — belongs to CNC, vacuum casting and bridge tooling. This page gives the four-way table, the tooling crossover math, and what changes on your drawing when you move between processes.
Compiled by the EKINSUN engineering team. Costs are typical planning ranges for small industrial parts — part size, cavities and finish move them. Last reviewed 18 Aug 2026.
| Your situation | Route | Why |
|---|---|---|
| 1 visual model, material doesn't matter | 3D printing (a print service) | Cheapest and fastest for a shape check — we honestly don't compete here |
| 1–50 functional parts, real material, threads, fits | CNC machining | No tooling, production material and tolerance from piece one |
| 10–200 plastic parts, real-part look and feel | Vacuum casting | Silicone mold from one machined master — production finish without steel |
| 200–5,000 plastic parts, design stable | Aluminium bridge tooling | $1,500–4,000 tool, molded-part unit price, weeks not months |
| 5,000+ parts, design frozen | Production injection molding | Steel tool amortizes to the lowest per-piece price |
| Metal part, any quantity with fits or strength | CNC (turning/milling) | Printed metal is a niche; machined metal is the production norm |
Three-way articles compare printing, machining and molding, then quietly skip the process that bridges them. The honest table has four columns:
| 3D printing | CNC machining | Vacuum casting | Injection molding | |
|---|---|---|---|---|
| Tooling cost | None | None | Silicone mold, low hundreds $ | $1,500–4,000 (aluminium) / $5,000–20,000+ (steel) |
| Sweet-spot quantity | 1–20 | 1–500 | 10–200 | 500+ (bridge) / 5,000+ (steel) |
| Per-piece price at its sweet spot | Low for shapes, high for quality | Mid | Between CNC and molding | Lowest — often $1–3 small parts |
| Materials | Printable resins/filaments; metals in special processes | Any machinable stock: 6061, 7075, 304/316, steel, POM, PA, PEEK | Production-like polyurethanes; Shore A to glass-filled rigid | The actual production polymer: ABS, PA6/PA66, POM, PP, PC |
| Typical tolerance | ±0.2–0.5 mm | ±0.05 mm standard, ±0.01 mm where specified | ±0.15–0.3 mm typical | ±0.05–0.2 mm, geometry-dependent |
| Threads & fits | Weak — printed threads wear | Full — cut threads, ISO 286 fits | Inserts molded in | Inserts or molded threads |
| Strength behaviour | Anisotropic (layer lines) | Isotropic — bar-stock properties | Near-production, cast PU | Production properties |
| Lead time, first parts | 1–3 days | 5–10 days | 1–2 weeks incl. master | 3–6 weeks (bridge) / 6–12 weeks (steel) |
| Design changes mid-run | Free | Free between batches | New mold, low cost | Mold rework or scrap |
The whole CNC-vs-molding decision compresses into one line: crossover quantity = mold cost ÷ (machined price per piece − molded price per piece). Worked example on a palm-sized ABS part: aluminium tool $2,500, CNC at $9/piece, molded at $1.20/piece → 2,500 ÷ 7.80 ≈ 320 pieces. Below that, machining spends less in total and keeps the design changeable; above it, every piece pays the mold back.
| Comparison | Typical crossover | Driver |
|---|---|---|
| CNC → vacuum casting | ~10–20 pieces | Master + silicone mold cost is small; casting cycles are cheap |
| CNC → aluminium bridge tool | ~300–1,000 pieces | $1,500–4,000 tool ÷ per-piece saving |
| Bridge tool → steel production mold | ~5,000–20,000 pieces | Steel tool cost and cavity count vs cycle life |
Two honest footnotes to the formula. First, it ignores design risk: a mold freezes the design, so if a revision is even possible, weight the crossover upward. Second, it ignores time: a machined batch ships while the mold is still being cut — some buyers order both, CNC parts now and molded parts at the crossover.
Most comparisons stop at "print small, mold big" — which strands the most common industrial order size we see: a few dozen to a few hundred pieces. That band has real answers, and they are exactly what we build daily:
EKINSUN LTD is a custom parts manufacturer in Guangdong, China. We machine, vacuum-cast and injection-mold — we do not run a print farm. For a shape-check prototype where the material is irrelevant, an online print service is the cheaper tool and the honest recommendation. The flip happens on the day the part has to work: real material grades, threads that survive service, ±0.05 mm features, or a batch beyond printing economics. From that day the routes on this page take over — CNC from 1 piece, vacuum casting from 10, molding from the crossover up — and your printed prototype, sketch or sample is a perfectly good starting document for any of them.
Not sure which band you're in? Send the part — file, sketch, photo or the printed prototype itself — and the quantity you actually need. We reply in 12 hours with the route, the price, and the crossover quantity where the next route becomes cheaper. Get a quote.
Divide the mold cost by the per-piece saving. Example: a $2,500 aluminium bridge tool, CNC at $9 per piece, molded at $1.20 per piece — 2,500 ÷ 7.80 ≈ 320 pieces is the crossover. For typical small plastic parts the crossover lands between roughly 300 and 1,000 pieces; below it CNC or vacuum casting wins on total spend, above it the mold pays for itself.
No — EKINSUN machines, vacuum-casts and injection-molds. For a visual-only prototype where material behaviour doesn't matter, an online print service is genuinely the cheaper tool and we say so. The moment the part must be the real material — machined threads, ±0.05 mm fits, production ABS/PA/POM, load-bearing metal — that is where our routes start, from 1 piece.
This is the band the three-way comparisons skip: too many pieces to print or machine economically, too few to justify a steel mold. Vacuum casting from one CNC-machined master fills it — a silicone mold costs a few hundred dollars, delivers about 20 shots per mold in production polyurethanes, and holds real-part finish. Above a few hundred pieces an aluminium bridge tool takes over until full production tooling makes sense.
Sometimes — but the limits arrive fast: layer anisotropy under load, weak or inaccurate threads, temperature limits of common printed resins, and surface finish that reads as a prototype. When a printed part starts failing in service or needs threads, fits or certified material grades, the production answer is CNC from solid stock or a molded part in the real polymer.
Typical ranges we quote against: aluminium bridge tooling $1,500–4,000 for simple single-cavity parts; production steel molds $5,000–20,000+ depending on cavities, sliders and texture. That upfront cost is the whole decision: it buys a per-piece price often below $1–3, but only quantity can pay it back — which is why the quote states the crossover quantity, not just the mold price.
On small plastic parts, usually not per piece — molding is faster per cycle. But CNC carries zero tooling, so below the crossover quantity the total spend is still lower, and CNC allows design changes between batches without scrapping a mold. At 1,000 pieces of a simple part the mold usually wins; at 1,000 pieces of a part still being revised, machining often still wins.
Four things: material becomes a real grade (ABS, PA6, POM, 6061 — not "resin"); internal corners gain radii for CNC cutters; molded versions need draft angles and roughly uniform wall thickness; and tolerances get re-stated per feature, because a printer's ±0.2 mm everywhere is neither achievable in molding nor necessary to pay for in machining. Send the printed part or its file and we return the production drawing with these changes marked.
Send the file, sketch or printed prototype + your quantity. We reply with the route, the price and the crossover point. MOQ 1, quote in 12 hours.
Machined, cast or molded — quoted with the route and the crossover quantity stated, from 1 piece. Reply in 12 hours.