CNC machining vs injection molding: the volume decision, made concrete

Guide · CNC vs Injection Molding CNC machining vs injection molding: the volume decision, made concrete CNC has no tooling and no minimum; injection molding has a steel mold and a per-part cost that collapses at volume. The right process is a function of quantity, tolerance and geometry — and the honest answer is often both. This guide gives you the decision framework, the break-even math, and the hybrid path that gets prototypes out in days and production costs down. How the two processes work CNC machining is subtractive: a cutting tool removes material from a solid block until only your part remains. Injection molding is formative: molten thermoplastic is forced into a closed steel cavity, cools, and is ejected as a finished part. The two consequences that drive every downstream decision are tooling and cycle time. CNC needs no mold — the first part and the ten-thousandth cost about the same to make, because cost is dominated by machine time. Injection molding needs a mold that costs real money and real weeks, but once that mold exists, each part is produced in seconds for a few cents of resin. Everything else — tolerance, material choice, surface finish, lead time, geometry — flows from that one structural difference. A useful way to hold the comparison: CNC buys speed and freedom with a higher unit cost; molding buys a low unit cost with upfront capital and a locked design. The rest of this guide walks the trade-offs in the order they matter to a sourcing decision. Cost structure: no tooling vs an amortized mold CNC cost is dominated by setup and machine time. A complex part costs more because it takes longer to cut, but there is no fixed investment you must recover before the first unit is profitable. Programming is a one-time cost measured in hours, and a fixture is a few hundred dollars, not tens of thousands. That is why CNC supports an MOQ of one: you can order a single machined part and pay only for that part's time. Injection molding cost splits into two clean buckets. The first is the mold — a precision steel assembly with cavities, cores, ejectors and cooling channels. A simple single-cavity prototype mold can run a few thousand dollars; a hardened multi-cavity production mold for a complex housing can run five figures and beyond. The second bucket is the part cost — resin plus machine time, which for a small part is often measured in cents. The mold is the toll you pay to enter production; the low part cost is the reward you collect only after enough parts spread that toll thin. The crossover formulaMold cost ÷ (CNC per-part cost − molded per-part cost) = break-even quantity. Worked example: a CNC part costs $18 at any quantity; the same part molded costs $1.40 after a $15,000 mold. The lines cross near 900 parts. But that is the naive break-even — the real number is higher, because you must also carry mold lead time, first-article revisions, and the risk of a design change that strands your tooling. Two hidden costs push the practical crossover well above the arithmetic one. First, design risk: if the part changes after the mold is cut, you pay for a new mold or a costly rework. CNC absorbs change for free. Second, cash flow and time: a mold ties up capital for weeks before a single part ships. A sourcing decision built on the formula alone — ignoring revision risk and time-to-market — is the most common way buyers over-tool too early. The volume break-even The crossover is not a single number; it scales with part size, complexity and material. Small, simple parts break even at lower volumes; large or highly machined parts stay competitive in CNC far longer. The table below is a first-pass screen, not a substitute for modeling your own numbers. Volume bandTypical outcomeWhy 1–100 partsCNC, every timeNo tool; mold cannot amortize 100–1,000 partsUsually CNCTooling still dominates total cost 1,000–10,000 partsModel itCrossover zone; design risk decides 10,000+ partsUsually moldingMold amortized; per-part cost wins Two qualifiers matter. A stable design in the crossover zone tilts toward molding; a design still in flux tilts toward CNC even at volumes where the formula says "tool it." And a deep CNC bench shifts the curve — a shop with 80+ machines and a one-part minimum can hold CNC cost flat and fast into volumes where a smaller shop would push you into a mold. Tolerance and repeatability This is the cleanest differentiator, and it usually decides the argument on its own. CNC machining holds ±0.005 mm as a routine tolerance, with ±0.002 mm available on precision grinding — and it does so with process control, not luck. SPC-monitored machining at Nex-G holds critical characteristics to Cpk ≥ 1.67, general features to Cpk ≥ 1.33 — accurate and repeatable. That is the tolerance envelope a bearing seat, a sealing face or a press-fit bore demands. Injection molding holds looser tolerances, typically ±0.1 mm to ±0.5 mm depending on part size and resin. Shrinkage is the root cause: plastic shrinks as it cools, and every resin shrinks differently — amorphous materials like ABS move less than semi-crystalline ones like polypropylene or nylon. A good molder compensates with gate placement and process control, but you are still buying tenths of a millimeter of capability, not microns. If a feature must be tight, molding alone will not get you there; you machine it, or you insert a machined component. The practical rule: specify molding where the design tolerates tenths, and reserve machining for the features that demand microns. If your drawing is full of ±0.05 mm callouts, CNC is the process — or a molded body with machined critical faces. Materials: machinable stock vs thermoplastics CNC is material-agnostic within reason. If it comes as bar, plate or billet and it is machinable, you can cut it: aluminum, stainless steel, brass, copper, titanium, magnesium and a wide range of engineering plastics as machined stock. That breadth is the point — a single CNC program can run aluminum today and stainless tomorrow, and you can prototype in a cheap material and switch to the production alloy without changing tooling. Full property data lives in the ASM Handbook. Injection molding is confined to thermoplastics (plus thermosets and liquid silicone rubber in specialist shops). The workhorses are ABS, polycarbonate, PC/ABS, nylon (PA), polypropylene, POM/acetal, and the high-temperature grades like PEEK and PPS for demanding environments. You cannot mold aluminum in this process — metals go to die casting or MIM, which are different animals with their own tooling economics. So the material question is often the first filter: if the part must be metal, molding is already off the table. There is a subtlety worth naming. Machined plastic stock is not the same as molded resin: a machined PEEK part is anisotropic, because stock has internal stress and grain, while a molded PEEK part is more uniform. For metal parts that must be strong, dense and weldable, machined billet also beats any net-shape process on mechanical properties. When the spec demands metal, CNC is not a compromise — it is the only correct route. Surface finish Molding wins on cosmetics out of the box. The cavity surface transfers directly to the part, so you can specify a glossy SPI A-grade finish, a matte texture, or a fine grain and get it on every shot with no secondary operation. Textures like soft-touch or leather grain are molded in, not applied later — a real advantage for consumer-facing housings. CNC leaves visible tool marks. A machined surface is honest about how it was made: you see the stepover of the cutter. That is fine for functional faces but rarely acceptable for a cosmetic exterior without post-processing. Getting a machined part to a premium finish means bead blasting, brushing, anodizing, or a secondary polish — all cost and all time. Where the part is functional or hidden, CNC's as-machined finish is irrelevant; where the part is the product's face, molding's molded-in finish is a genuine cost advantage. Lead time: days vs weeks This is where CNC dominates and the reason it anchors every development program. A machined part ships in days. With no tool to build, first parts can be quoted, programmed and cut on lead tiers of roughly 3 / 7 / 30 days depending on complexity. A prototype can be in your hands the same week you release the drawing. Injection molding waits on the mold. Tool design, steel, cavity cutting, fitting and first-article trials commonly run four to eight weeks — and longer if the first shots reveal a fill or shrink problem that sends the tool back for rework. You cannot compress that easily, because the mold is a precision mechanical assembly, not a file you print. The consequence for a hardware program: you cannot start production on a molded part until the tool is proven, but you can start a machined part the moment the design is frozen — or even before. That lead-time gap is what makes the hybrid so powerful: machine the first units now to feed EVT and DVT, and let the mold build in parallel so it is ready for PVT and production ramp. The math is simple — every week saved to first article is a week of development velocity, and for a hardware program that week is often worth more than the difference in unit cost. Part geometry: undercuts, thin walls, threads Geometry pushes the decision in both directions, feature by feature. Thin walls favor molding. A molded wall of 1–2 mm is routine and nearly free — the resin flows into it in milliseconds. A machined thin wall is the opposite: it deflects and chatters under the cutter, and holding it flat and true costs disproportionate machine time. If the part is a thin-walled enclosure, molding is built for it. Undercuts favor CNC. A machined undercut needs only the right toolpath or a 5-axis setup. A molded undercut needs a side action, slide or lifter — mechanism inside the mold that raises its cost and complexity, and sometimes cannot be done at all without a design change. Every undercut you add to a molded part is money you add to the tool. Threads split the difference. CNC thread-mills internal and external threads directly into the part, no inserts required. Molding threads requires either an unscrewing core (expensive, slow) or a design that accepts a molded boss later tapped or a heat-staked insert. The cheap production answer is usually: mold the boss, machine or stake the thread. The throughline: molding loves organic, thin, net-shape geometry; CNC loves features that must be cut, located or held tight. Most real parts contain both, which is why the two processes are so often sequenced rather than chosen. Where each process wins CNC wins when Volume is low or uncertain, and MOQ is one The design is still changing Tolerances are tight (±0.005 mm and finer) The material must be metal You need parts in days, not weeks Geometry has deep pockets, undercuts or internal threads Injection molding wins when Volume is high and the design is stable The part is a thin-walled plastic enclosure Cosmetic finish is molded-in Per-part cost must be minimal You need many identical parts, fast, once the tool exists The trap to avoid is treating these as rival religions. They are tools with different cost curves, and the winning answer for a real product is usually a sequence: machine now, mold later. The hybrid: prototype in CNC, produce in molding The dominant real-world pattern is not either-or — it is a handoff. You machine prototypes and pilot runs in CNC while the design is still settling, because machining absorbs every change for free and ships in days. That keeps EVT and DVT moving through the EVT → DVT → PVT gates without waiting on a tool. When the design freezes and volume is proven, you cut the mold, validate first articles, and transition to molded production — with the machined parts having already de-risked the fit, the function and the market. The hybrid also runs the other direction at production. Even when a part is molded, the features that matter — sealing faces, bearing bores, threaded bosses, datum surfaces — are frequently machined after molding to hit tolerances the resin cannot. Molded body, machined critical features: that is how a plastic part holds a real seat. A partner that runs CNC in-house and molding through vetted partners can execute the whole sequence under one quality system instead of forcing you to manage two suppliers across a handoff. Buyer decision table Map your part to a process. "Hybrid" means machine now for development, mold for steady-state volume, and machine the critical features on molded parts. Decision factorCNC machiningInjection moldingHybrid ToolingNoneSteel mold, $k–$$k, 4–8 wkMold + machining setup Minimum orderOne (MOQ 1)Thousands to pay the toolOne now, volume later First-part lead timeDays (3–30 by complexity)Weeks (mold first)Days now, mold in parallel Per-part costFlat, moderateLow at volumeLow body + targeted machining Tolerance±0.005 mm±0.1–0.5 mmMolded body, machined faces MaterialsMetals + engineering plasticsThermoplasticsPlastic body, metal inserts Surface finishTool marks; needs finishingMolded-in, SPI/textureMolded cosmetics, machined seats Design changeFreeCostly mold revisionChange before tooling, not after Best forPrototypes, spares, metal, tightStable high-volume plasticReal product lifecycles Decision checklist: CNC or molding Before you model the break-even, run this five-question checklist. It sorts most parts into the right process in about a minute, and it catches the two mistakes that cost the most: tooling too early, and machining too long. Is the design frozen? If the geometry is still moving, CNC — a mold locks the shape, and every change after tooling is a paid revision. Is the material metal? If yes, molding is off the table — CNC (or die casting/MIM, which carry their own tooling) is the route. Are there micron-level features? If a bore, seat, or sealing face must hold ±0.05 mm or tighter, plan to machine it — either the whole part or that feature on a molded body. What is the honest volume? Under a thousand parts, CNC usually wins outright; above ten thousand, molding usually does. Between them, the crossover is decided by design risk and time-to-market, not the formula alone. When do you need first parts? Days favors CNC; weeks favors molding. If the program needs parts now and volume later, the answer is both — machine now, mold in parallel. Answer all five, and the process choice is usually obvious. Where it is not, the hybrid closes the gap: machine the development units, tool the mold when the design and volume are proven, and machine the critical features on the molded parts that follow. Nex-G: CNC in-house, molding via vetted partners Nex-G runs the machining side of this decision in-house and the molding side through vetted partners, so you get one partner across the crossover instead of two suppliers across a handoff. From our Dongguan Hengli plant — 6,800 m², 100+ staff, operating since 2006 — we deliver EMS assembly and CNC machining on 80+ machines from Mazak, Brother, TSUGAMI and Sodick, holding ±0.005 mm (and ±0.002 mm on precision grinding) with SPC control to Cpk ≥ 1.67. We carry ISO 9001, IATF 16949 (design excluded, clause 8.3) and ISO 14001, with URS audit to 2027. MOQ is one part, with lead tiers of roughly 3 / 7 / 30 days — so a prototype can be in your hands the same week the drawing is released. For plastic production, injection molding is sourced through qualified partners whose tooling we specify and whose first articles we validate against the machined parts we already made. That means the machined prototype and the molded production part are held to the same standard, by the same quality system, from the same Dongguan team — no disconnected suppliers, no lost handoff. Machine the first hundred now, tool the mold when volume proves out, and machine the critical faces when the molded parts land. That is the whole framework in one engagement. Frequently asked questions When is injection molding cheaper than CNC?Above the break-even volume, where the mold cost is amortized across enough parts that the low per-part resin cost beats flat CNC machine time. For a small part that crossover is often in the low thousands, but design risk and lead time push the practical number higher. Can injection molding hold tight tolerance?Not to CNC levels. Molding typically holds ±0.1–0.5 mm because of resin shrinkage. Critical faces are machined after molding, or inserts are used, when microns are required. Which is faster to first part?CNC by a wide margin — days, with no tool. Injection molding waits on the mold, commonly four to eight weeks before a first article ships. What if my design is still changing?Choose CNC. A mold locks the geometry; change it after tooling and you pay for a revision or a new mold. CNC absorbs change for free, which is why it anchors EVT and DVT. Can I start in CNC and move to molding later?Yes — that is the standard bridge. Machine prototypes and pilot runs while the design settles, cut the mold once volume proves out, and machine the critical features on the molded parts. What materials can each process use?CNC cuts metals and engineering plastics from stock. Injection molding is limited to thermoplastics such as ABS, PC, nylon, POM and PEEK. If the part must be metal, molding is off the table. Do you do both processes?We machine in-house on 80+ CNC machines and source injection molding through vetted partners, so you get one quality system across the crossover — machined prototypes, then molded production, then machined critical faces. What is the minimum order?For CNC, MOQ is one part. For injection molding there is no hard minimum, but a mold only pays for itself across thousands of shots, so tooling rarely makes sense below the crossover volume. What is the cheapest way to bridge CNC and molding?Machine the first units in CNC while the design settles, cut the mold only when volume proves out, then machine the critical features on the molded parts. That sequence de-risks the fit and the market before a dollar goes into tooling. Can I switch from molded back to CNC later?Yes — CNC has no minimum and no tool, so it can cover spares, low-volume variants, or a design refresh after a mold is retired. Keeping the CNC program alive alongside the mold is cheap insurance. Not sure where your part crosses over?Send the drawing, material and volume to [email protected] — we will model the break-even and recommend CNC, molding, or the hybrid that costs least.Request a quote Related articlesDie Casting vs CNC MachiningPlastics Machining GuideCNC Machining Cost in China