China manufacturing MOQ: why no-MOQ changes the math

MOQ · China Manufacturing China manufacturing MOQ: why no-MOQ changes the math Minimum order quantity is the silent tax on every hardware program — it forces you to over-buy at stages where you should be learning. This guide explains what MOQ is, why factories impose it, when it matters and when it does not, and how a no-MOQ partner reshapes your whole cost and risk profile. Request a quoteWhat MOQ is What MOQ actually is Minimum order quantity is the smallest batch a supplier will accept. It exists for a reason: setup, tooling and material handling have a fixed cost that only makes sense above a volume. For processes with hard tooling — injection molding, die casting, stamping — the mold cost must be amortized, so a high MOQ is structural. For processes without tooling — CNC machining, PCB assembly — there is no physical reason for a minimum, and an MOQ there is a business policy, not a technical one. Where tooling forces MOQ. Why factories set a minimum: setup, tooling and component reels For a manufacturing engineer. A minimum is always the sum of three fixed costs. Understand them and you can tell at a glance whether an MOQ is structural or simply the supplier's preference. Setup Every job starts the same way: machine programming, tool selection and fixturing. On a CNC mill the spindle has to be programmed and the work held; on an SMT line the feeder positions have to be loaded and the pick-and-place program written. That setup time is identical whether you order one piece or one thousand, so a factory that charges a setup fee must spread it across a batch to make the job worth taking. A shop that has built its lines for quick changeovers absorbs that cost instead of passing it back to you as a minimum. Tooling Injection molds, die-cast dies and stamping tools are cut steel that only pays for itself across tens or hundreds of thousands of shots. This is why a molded part carries a hard minimum: the supplier is not being difficult, it is protecting the amortization of a tool it built on your behalf. If the design later changes, that tool is scrap — which is exactly why tooled processes and unproven designs are a bad combination. Component reels Surface-mount assembly pulls parts from reels, and a standard reel holds thousands of identical parts. If your bill of materials calls for 12 of a component, the line still has to load a reel that costs far more than the 12 parts you need, and the leftover sits on a shelf. No-MOQ shops solve this with partial-reel handling and consolidated purchasing across customers, but the reel is the real reason an SMT minimum exists in the first place. Why MOQ is a tax on your program Hardware volumes change at every stage of NPI: five parts at EVT, fifty at DVT, five hundred at PVT, ten thousand at production. A supplier with a 1,000-piece minimum forces you to over-buy at the early stages — you pay for a thousand when you need five, and the surplus sits as stranded inventory or gets scrapped when the design changes. The cost is not just the cash; it is the lost learning, because over-buying removes the incentive to iterate. Why NPI volumes change every stage. The true cost of a low MOQ For a sourcing engineer. Low volume costs more per unit, and that is not a penalty — it is arithmetic. Setup, programming, inspection and the first article all happen once whether you buy one piece or one hundred, so the unit price falls as the batch grows. A buyer who chases the lowest unit price at prototype stage is optimizing the wrong variable: at five pieces the dominant cost is not the part, it is the time and risk of the iteration. What a low-MOQ price buys is the removal of waste, not the removal of cost. You pay a higher unit price so you do not pay for 995 parts you will scrap when the revision lands. The comparison that matters is total program spend across the whole NPI cycle, not the line item on one purchase order. Measure cost per shipped product, and low-MOQ usually wins — because it never forces you to carry inventory against a design you have not yet proven. There is also a hidden cost on the other side that never appears on a quote: the carrying cost of surplus. Over-bought parts occupy shelf space, tie up working capital and accrue obsolescence risk while your design is still moving. When you net those out against a higher early unit price, the low-MOQ path is often cheaper in the only number that reaches the P&L: cash actually consumed before the product ships. When MOQ genuinely matters High-volume molded or cast parts. The tool cost is real; amortizing it across volume is rational. Commodity components bought at scale. Suppliers set minimums that reflect their own lot sizes. Even here, the question is whether you are at the volume that justifies the tool — and whether the design is frozen enough that the tool will not be stranded. When MOQ does not matter (and should not exist) CNC machining. No tool; order one or one thousand. PCBA assembly. No tool; the constraint is components, not a minimum batch. Low and medium volume electronics. The whole point is flexible, small-batch build. For these, a no-MOQ partner is not a discount — it is the correct operating model for a product still finding its shape. Prototype-to-production scaling: 3 / 7 / 30 days A no-MOQ shop should quote lead time by complexity, not by volume. The standard is three tiers: a simple prototype in about 3 days, a standard small batch in about 7 days, and a complex production part in about 30 days. The spread is driven by process steps — a part that moves through 10 to 20 operations, especially one that needs finishing or surface treatment, lands at the long end. That is honest lead time: the number is set by the work, not by how many you order. 1piece minimum — no MOQ, from prototype to MP 3 / 7 / 30days: prototype / small batch / complex production 80+CNC machines on one Dongguan floor ±0.005mm machining tolerance (±0.002 grinding) This lets you plan the ramp without guesswork. A single machined bracket for an EVT build is a 3-day item; a batch of 50 for DVT with finishing is a 7-to-30-day item; the full production run inherits whatever the line needs once the design is frozen. One Dongguan facility — 6,800 m² in Hengli, 100+ staff, established 2006 — carries a one-off prototype and a thousand-piece run through the same 80+ machining centers and lathes from Mazak, Brother, Citizen and Tsugami, all under one quality system, so the parts you validate at DVT are the parts you ship at MP. First-article inspection on every run Small runs do not mean relaxed inspection. Every job starts with a first article measured on a Hexagon CMM or optical inspection equipment, and every part is traceable through MES/ERP — material, process parameters, equipment, batch, operator and inspection data are logged. A 50-piece DVT batch gets the same dimensional evidence as a 10,000-piece production run, because the number of pieces does not change the number of checks that matter. The no-MOQ model A no-MOQ shop built for NPI lets you order exactly what each gate needs: five for EVT, fifty for DVT, and only scaling to volume when the design is proven. The learning stays cheap, the inventory stays low, and the scaling is deliberate rather than forced. It also removes the stranded-tooling risk entirely — there is no mold to abandon if the design changes. The EMS build path. NPI gates: how no-MOQ maps to EVT, DVT, PVT For a program manager. A no-MOQ partner maps one-for-one onto the three gates. At EVT you order a handful to prove the function; at DVT you order production-intent parts to prove the design survives real conditions; at PVT you run a low-hundreds batch on the same line that carries mass production, so the capability data you sign off is real, not theoretical. GateTypical volumeWhat you orderEvidence you get EVT5 – 50A handful to prove functionBuild log, corrected BOM DVT50 – low hundredsProduction-intent partsDimensional report, locked tolerances PVTLow hundreds – ~1,000Line-built batchControl plan, SPC / Cpk data The gates also tell you what evidence to demand. At EVT you want a build log and a corrected BOM. At DVT you want a dimensional report and locked tolerances. At PVT you want the control plan and process capability data. The automotive work shows what that evidence looks like in practice: 67 threshold for a key characteristic. That is the difference between a line that says it holds spec and one that can prove it. The three gates in detail. How no-MOQ changes risk for startups For a startup, a high MOQ converts engineering risk into cash risk. You are betting capital that the design will not change — and at EVT the design always changes. Every piece you over-buy is inventory you must store, count and eventually write down when the revision lands. No-MOQ inverts that. It caps cash exposure at exactly what the next gate needs, so a wrong assumption costs a redesign instead of a write-off. It removes stranded-tooling risk, because CNC and PCBA need no mold — there is nothing to abandon if the geometry changes. And it preserves the cheapest iteration window in the product lifecycle: the time before you have committed volume. A supplier that lets you order five at a time is the only kind that lets you exploit it. How to find and use a no-MOQ partner Ask directly. “What is your minimum order?” A confident “one” tells you they are built for it. Check the process range. If they own both EMS and CNC, one partner carries you from EVT to MP without a handoff. Verify they scale. No-MOQ at prototype is common; no-MOQ at production with stable cost is the real test. Plan the volume ramp. Use the freedom to learn cheaply, then commit volume when the design is frozen. Minimum-order economics: why a minimum is a business decision A minimum is a statement about the supplier's cost structure, not your product. A shop that runs long, uninterrupted production of a few part numbers wants a high MOQ because changeovers destroy its efficiency. A shop built for NPI — quick programming, short setups, flexible lines — makes money on small runs, so it has no reason to impose one. That is why the same part can carry a 1,000-piece minimum at one supplier and no minimum at another: the difference is the factory's operating model, not the physics of the part. For CNC machining and PCB assembly there is no tool to amortize, so a minimum there is pure policy. When you see an MOQ on a no-tool process, you are seeing a supplier optimized for volume pricing you out of the early stages — and that is your cue to find a partner optimized for NPI instead. The EMS and CNC split The two no-tool processes — electronics assembly and machining — differ in where the minimum hides. On CNC the floor is a fixture and a program; the marginal piece is cheap, so a one-piece minimum is natural. On SMT the component reel is the constraint, so the practical minimum is governed by your bill of materials, not the line. A shop that runs both under one roof can quote the honest minimum for each and let one purchase order carry the enclosure and the board together. MOQ vs cost: the real trade-off ScenarioWith MOQWith no-MOQ EVT (need 5)Buy 1,000, scrap 995 on changeBuy 5, iterate freely DVT (need 50)Still bound by 1,000 minimumBuy 50, validate MP (need 10,000)MOQ finally metScale to 10,000 at stable cost The no-MOQ path costs more per unit early only because you are not amortizing a forced batch — and you avoid paying for parts you never use. The buyer's low-volume checklist Ask the minimum outright. A confident “one piece” is the signal you want; a hedge is not. Ask for lead time by complexity. Three tiers — about 3 / 7 / 30 days — set by process steps rather than order quantity are a healthy answer. Confirm one line, one quality system. The line and controls that run your prototype should be the ones that run production, so DVT capability data transfers to MP. Ask for capability, not a certificate. A real Cpk on a real dimension beats a wall of logos. Check the process range. One partner that owns EMS — SMT, box build, testing — and CNC carries you from EVT to MP without a handoff. Price the program, not the PO. Compare total spend across the NPI cycle, not the unit line on a single quote. Worked example: what a forced MOQ costs across a program Numbers make the trade-off concrete. Say a machined bracket costs $8 each at a 1,000-piece minimum, or $26 each ordered five at a time with no minimum. Buy the thousand at EVT and you tie up $8,000 before the first test is done — then the second revision moves a mounting hole and 995 of them become scrap. Order five at a time and the EVT build costs $130. Carry the same bracket through DVT (50 units) and you are at $1,300 on the no-MOQ path, against an $8,000 minimum order you still cannot use. By the time the design is frozen at PVT and you genuinely need 1,000, the no-MOQ partner is quoting the same $8 unit cost — and you arrive there having spent under $2,000 to learn instead of more than $16,000 across two forced batches. The point is not that no-MOQ is cheaper per unit; it is that the minimum orders you are forced into early are the most expensive parts you will ever buy. No-MOQ is not the same as no inventory plan A no-MOQ partner removes the supplier's minimum, but it does not remove the component-level minimums baked into your BOM — reels, trays, tubes and a distributor's own lot sizes still set a floor on what you can buy. The difference is where the risk sits: with no-MOQ you are never forced to buy finished goods you cannot use, you only absorb the component lots your bill of materials genuinely needs. For steady-demand products, pair no-MOQ with a safety-stock or vendor-managed inventory agreement so reorder lead time does not become your ceiling. No-MOQ removes forced finished-goods over-buy; a deliberate inventory plan removes stock-outs — they solve different problems, and a mature program needs both. On the floor the checks are concrete rather than aspirational. Every BOM line is cross-checked against the drawing and the purchase spec before release, so a component callout cannot silently change between EVT and PVT. Critical parts and alloys are bought through a cross-checked second source with a matching certificate, and each incoming lot is confirmed before it reaches a line. For programs that require it we run FAI per AS9102 so the first article is signed off before the batch begins, under an AS9100-aligned quality system. Frequently asked questions What does MOQ mean in manufacturing?Minimum order quantity — the smallest batch a supplier will accept. For tooled processes (molding, casting) it reflects real tool amortization; for CNC and PCBA it is usually a policy, because there is no tool to amortize. Why is MOQ a problem for new products?Because NPI volumes change every stage (5 at EVT, 50 at DVT, thousands at MP). A high minimum forces over-buying at early stages, stranding inventory and removing the incentive to iterate when learning is cheapest. Which processes have no real MOQ?CNC machining and PCB assembly have no tool, so there is no physical reason for a minimum — you can order one or one thousand. High-volume molding and casting are the opposite, because the tool cost must be amortized. Is no-MOQ more expensive per part?Early on, yes per unit — because you are not forced to amortize a large batch. But you avoid paying for parts you never use, and you keep the freedom to change the design, which is worth more than the unit saving at prototype stage. Do you have a minimum order?No. Nex-G runs no-MOQ from prototype SMT through mass production, so you order exactly what each NPI gate needs and scale when the design is proven — no stranded tooling, no over-buy. How do I know a shop is truly no-MOQ?Ask the minimum directly and check they own the process range (EMS and CNC) so one partner carries you without a handoff. The real test is stable cost at production, not just at prototype. What lead times can I expect at low volume?Three tiers by complexity: about 3 days for a simple prototype, about 7 days for a standard small batch, and about 30 days for a complex part spanning 10–20 process steps or needing finishing. The number is set by the work, not by the order quantity. Do your CNC and EMS share one facility?Yes. A single 6,800 m² site in Hengli, Dongguan — 100+ staff, established 2006 — runs EMS (SMT, box build, testing) and CNC under one roof and one quality system, so prototype and production parts come from the same line. What quality evidence do you provide?We build automotive-grade APQP/PPAP packages. Certifications: ISO 9001, IATF 16949 and ISO 14001, all URS-registered. Does IATF 16949 mean you take design responsibility?No. Our IATF 16949 certificate is scoped to the manufacture of CNC metal parts and explicitly excludes product design under Clause 8.3. We build to your drawing and your tolerances; we do not own the design. How does no-MOQ work with SMT component reels?A reel can hold thousands of parts while your BOM needs a handful, so the minimum on a component line is real. We handle it with partial reels and consolidated purchasing, so that constraint does not become your over-buy. What tolerance can you hold?±0.005 mm on CNC milling and turning, and ±0.002 mm on precision grinding, verified on Hexagon CMMs and optical inspection equipment in-house. Can I order a machined part and a small PCBA run in one order?Yes. Because EMS and CNC sit under one roof, a single purchase order can carry the enclosure, brackets and the assembled board — which removes a handoff and the minimums that come with splitting suppliers across two factories. Does a low MOQ mean lower quality?No — the number of pieces does not change the number of checks that matter. A one-piece order still gets a first-article inspection on a Hexagon CMM and full material traceability, because the process is controlled by measured capability and documented control plans, not by batch size. Building something still changing?Email your stage and volume to [email protected] — we will quote exactly what you need now, with a clear path to scale when the design is frozen.Request a quote Related articlesNew Product Introduction (NPI)The Complete Guide to Electronics ManufacturingDie Casting vs CNC MachiningManufacturing in Dongguan