Key takeaways
- Where to manufacture is four nested decisions, not one: what the product is made of, year-one volume, geography, and how many vendors you will manage.
- The honest volume number changes everything downstream. Budgeting tooling for 10,000 units while planning for 500 is the classic error.
- Manufacturing choices belong in the design phase, not after it, because each one locks before the money is actually spent.
- Some costs only appear once you commit: tooling amortisation against unsold volume, customs and processing fees, and the merchandise processing fee.
Where to manufacture your product comes down to four separate decisions, made in a specific order: what the product is made of, how many units you actually need in year one, which geography fits your volume and risk tolerance, and how many vendors you are willing to manage. Get the order wrong and you will lock in a geography before you know your real volume, or commit to tooling before your design is stable. This guide walks through all four, in the order they should be decided, plus the costs and failure points that show up only after you commit.
The decision is really four decisions, not one
Founders often treat “where should I manufacture” as a single choice between, say, China and the United States. In practice it is four nested decisions, and each one constrains the next.
- Material and process. What your product is made of rules out entire categories of manufacturer before location even matters.
- Volume. Your honest year-one unit count determines which tooling and which manufacturers are viable at all.
- Geography. Offshore, domestic, and nearshore each trade cost against lead time and control differently.
- Vendor count. One vertically integrated partner versus several specialists is a management decision as much as a cost one.
Decision one: what your product is made of, and what that rules out
Plastics, metal, and electronics have different rules
A plastic housing, a machined metal bracket, and a populated circuit board are built by different kinds of manufacturers, using different tooling logic, and the volume at which each process becomes economical is different too.
| Process | Typical use | Where it becomes cost-effective |
|---|---|---|
| 3D printing | Prototypes, very low volume, complex geometry | Single units through the low tens |
| CNC machining | Metal parts, precision components, bridge production | Low tens through low hundreds of units |
| Injection molding (aluminum/bridge tooling) | Pre-production and low-volume plastic parts | Roughly 500 to a few thousand units |
| Injection molding (steel tooling) | Full production plastic parts | Several thousand units and up, where tooling cost amortizes |
| PCB assembly (PCBA) | Populated circuit boards | Viable from low tens of units for simple boards; complexity raises the practical minimum |
These figures are directional, not fixed. Every manufacturer publishes its own minimum order quantity (MOQ), and we found no regulatory or standards-body figure that defines a single crossover point between these processes. Treat the table as a starting map for a conversation with a specific vendor, not a guarantee.
Electronics carry an extra layer most founders underestimate: a populated circuit board is only as good as the assembly quality behind it, and assembly quality is graded against defined acceptance criteria rather than left to a factory’s own judgment. The electronics manufacturing industry uses acceptability standards published by the Global Electronics Association (formerly known as IPC, and the standards still carry their IPC designations, such as IPC-A-610 for acceptability of electronic assemblies). Ask any electronics manufacturer which class of assembly they build to and how they document it, since “PCBA” alone tells you nothing about the acceptance bar a board is held to.
Rapid prototyping methods also matter here, not just for early proofs of concept but for deciding when to lock a design. Our guide to how 3D printing changed prototyping, and what it did not change about manufacturing covers where printed parts help and where they mislead founders about how a molded or machined part will actually behave.
Decision two: how many units you actually need in year one
Why the honest volume number changes everything downstream
A founder who plans for 500 units but budgets tooling as though 10,000 units were certain has essentially pre-paid for volume that may never happen. Steel tooling only pays for itself once its higher upfront cost is amortized across enough units. If you are not confident in your year-one number, aluminum or bridge tooling, at a higher per-unit cost, keeps your capital exposure lower while you validate demand.
- Under-estimate your volume, and you pay a premium per unit indefinitely because you tooled for a smaller run.
- Over-estimate it, and you pay for steel tooling and MOQ commitments you cannot sell through.
- A staged approach, aluminum tooling for the first production run and a re-tool to steel once volume is proven, is a legitimate way to manage this uncertainty rather than a compromise.
Decision three: geography
Offshore, domestic, and nearshore in one page
Offshore manufacturing, historically concentrated in China, typically offers the lowest per-unit cost at volume but carries longer lead times and added import costs. Nearshore manufacturing, largely Mexico for US-based brands, offers shorter transit times and can qualify for reduced tariff treatment under the USMCA trade agreement, at a labor-cost premium relative to China. Domestic manufacturing carries the highest per-unit labor cost but the shortest lead time, the easiest in-person oversight, and no import duties at all.
| Offshore (e.g. China) | Nearshore (e.g. Mexico) | Domestic (US) | |
|---|---|---|---|
| Typical transit time | Weeks by ocean freight | Days by truck | Days |
| Per-unit cost at volume | Generally lowest | Mid-range | Generally highest |
| Import duties | Federal tariffs apply and are currently a significant, changing cost; verify current rates against the USTR’s Section 301 tariff actions before budgeting | Can qualify for reduced or zero tariffs under USMCA rules of origin | None |
| Oversight | Harder, more time zones and distance | Easier than offshore, still requires travel | Easiest |
Landed cost, the true delivered cost per unit, also includes US Customs and Border Protection processing fees on imports. Under federal customs regulations, the merchandise processing fee is an ad valorem 0.3464% of the shipment’s value, subject to an annually adjusted minimum and maximum per entry, and a separate harbor maintenance fee of 0.125% applies to cargo moving through qualifying US ports under 19 CFR § 24.24. These federal fees are stable and predictable. Tariff rates are not: they have changed multiple times in 2026 and should be verified at the time you are budgeting a specific shipment, not assumed from an article written months earlier.
Decision four: how many separate vendors you are willing to manage
You can work with a single vertically integrated partner that handles design, engineering, and manufacturing coordination under one roof, or you can assemble a team of specialists: an industrial design studio, an engineering firm, and a separate contract manufacturer or EMS provider. The trade-off is control and specialization against coordination overhead. Every handoff between vendors is a place where a requirement can get lost or a design assumption can fail to transfer. Our guide to what a vertically integrated product development company actually does covers this trade-off, including the evidence on why handoffs between vendors create risk, in more depth.
The vendor types themselves have specific meanings worth knowing before you shortlist anyone. An OEM in the US procurement sense is the brand that owns the design; an ODM already owns a working design it adapts for your brand; an EMS company builds to your finished documentation but does not design the product. Our full comparison of OEM vs ODM vs EMS walks through which type fits which stage.
When each decision has to be locked
Why manufacturing choices belong in the design phase, not after it
A popular claim in manufacturing marketing content holds that 70 to 80 percent of a product’s manufacturing cost is locked in at the design stage. A 1993 MIT Sloan School of Management working paper by Karl Ulrich and Scott Pearson traced that figure to informal, non-rigorous industry surveys rather than controlled data, and their own study of a set of consumer products found cost variation attributable to design choices was not clearly larger than variation attributable to differences in the manufacturing system itself. The safer, defensible version of the claim is directional, not precise: manufacturing decisions made early, before tooling is committed, are far cheaper to change than the same decisions made after a mold has been cut. Design for manufacturing (DFM) review should happen before tooling starts, not after a production validation failure reveals the design cannot be built at the volume or tolerance you need.
What a realistic manufacturing chain looks like end to end
- Design freeze and DFM review. The design is locked and reviewed specifically for how it will be built, not just how it will function.
- Sourcing and quoting. Multiple manufacturers quote against the same finished documentation, so the comparison is apples to apples rather than each vendor pricing a different interpretation of the design.
- Tooling. Molds, fixtures, and jigs are built or procured, the single largest fixed cost most founders underestimate. Tooling typically takes weeks, not days, and any late design change after tooling starts costs real time and money to correct.
- First article inspection. The first units off the tooling are checked against the design before a full run is approved.
- Pilot or bridge run. A smaller run, often on interim tooling, validates the process at production intent before the full commitment.
- Production run. Units are built at the committed volume, with quality checks at defined intervals rather than only at the end of the line.
- Logistics and import. Freight, customs clearance, and any applicable duties move the product from factory to your warehouse or fulfillment partner.
How long this chain takes, start to finish, is a separate question from where it happens, and it varies significantly by product complexity and certification requirements. Our guide to how long it takes to develop a hardware product breaks down where the calendar time actually goes, and our full hardware development cost breakdown covers the budget side stage by stage.
The costs that only appear once you commit
- Tooling amortization mismatch. You paid for steel tooling sized for a volume you have not sold yet.
- Customs and processing fees. The merchandise processing fee and harbor maintenance fee described above apply on top of the unit price and freight quote you negotiated.
- Inventory carrying cost. Longer offshore lead times mean holding more safety stock, which ties up cash even when the per-unit price looks lower.
- Requalification cost. Changing factories after production has started often means re-running first article inspection and, for regulated products, recertification.
Common ways this decision goes wrong
- Committing to a geography before the design is stable, then paying to requalify tooling after a late design change.
- Choosing a single low-cost vendor with no visibility into what is subcontracted, then discovering a critical process step happens somewhere you never vetted.
- Sizing tooling for an optimistic volume forecast instead of a conservative one.
- Treating the tariff and landed-cost picture as fixed, when it is one of the more volatile inputs in the entire budget and needs to be re-checked close to the time you commit.
- Assuming a lower factory quote is automatically a lower total cost, without adding freight, customs fees, current tariffs, and the inventory you will need to hold to cover a longer lead time.
- Skipping first article inspection on a repeat order because “the factory already knows how to build it,” and discovering a process drift only after a full run ships.
The volatility in landed cost is worth restating plainly: the tariff figures in the geography table above reflect a rate structure that has already changed more than once in 2026, following a Supreme Court ruling that voided an earlier set of tariffs and a subsequent replacement regime. Whatever rate you find quoted anywhere, including here, verify it against current USTR and CBP guidance at the time you are actually committing to a shipment, not at the time you first researched the decision.
How to weigh total cost of ownership, not just the quote
The number a factory quotes you is a per-unit price, not your total cost. Two quotes that look identical on paper can produce very different total costs once you account for tooling amortization, freight mode, customs fees, tariff exposure, and how much inventory you need to hold to cover the lead time. A slightly higher per-unit price from a domestic or nearshore vendor, with a shorter lead time and lower safety-stock requirement, can beat a lower offshore quote on a total-cost basis, particularly at lower volumes where you cannot yet absorb a long replenishment cycle if you sell out.
- Freight mode matters as much as freight cost. Ocean freight is cheaper per unit but takes weeks; air freight is faster but expensive enough to use only for launch inventory or emergency restocks.
- Quality variance has a cost even when it does not show up on an invoice. A factory with inconsistent yield forces you to over-order to cover defective units, which is a hidden cost against your true per-unit price.
- Currency exposure is real for offshore sourcing. A quote denominated in a foreign currency can move against you between the time you place an order and the time you pay for it.
A working checklist before you commit to any factory
- Confirm the material and process fit your design, not the other way around.
- Write down a conservative year-one volume number, and size tooling to it, not to your hoped-for number.
- Get a landed cost estimate that includes freight, customs processing fees, and current tariff rates, not just the factory’s per-unit quote.
- Decide how many vendors you are willing to coordinate, and weigh that against the cost of a single integrated partner.
- Complete a DFM review before tooling starts, not after a failed validation run reveals a problem.
- Verify the factory’s actual capability, not just its sales claims. See our guide to finding the right manufacturer for a hardware product for the specific verification steps.
Frequently asked questions
Is offshore manufacturing always cheaper than domestic?
Usually cheaper on a per-unit basis at volume, but not always cheaper on a landed-cost basis once freight, customs fees, current tariffs, and inventory carrying costs are included. The gap has narrowed as tariff costs on Chinese-origin goods have increased.
How do I know if my volume justifies steel tooling?
There is no universal threshold; it depends on your specific part geometry and the tooling quote you receive. As a starting framework, aluminum or bridge tooling keeps capital exposure lower below roughly a few thousand units, with steel tooling paying for itself as volume climbs into the low thousands and beyond.
Can I switch factories after production has started?
Yes, but expect to repeat first article inspection and, for regulated products, some recertification. Budget both the direct cost and the schedule delay before assuming a switch is simple.
Does nearshore manufacturing avoid tariffs entirely?
Not automatically. Products manufactured in Mexico can qualify for reduced or zero tariffs under USMCA rules of origin, but qualification depends on where the components and value-added work actually happen, not just the country of final assembly.
What is the biggest mistake founders make in this decision?
Locking in a geography and tooling commitment before the design is stable and before the year-one volume number is realistic, then paying twice: once to build it wrong, and again to requalify it correctly.
Should I manage manufacturing myself or use an integrated partner?
That depends on how much coordination overhead you can absorb and how confident you are vetting each vendor independently. Our comparison of how to choose a hardware product development partner walks through that decision directly.
Where Inventornest fits
Inventornest coordinates material selection, tooling, and manufacturing location decisions when the agreed scope includes manufacturing preparation, so the DFM review happens before tooling is committed rather than after a production run reveals a problem. If you want a second opinion on a quote you already have, you can get a quote. Our OEM development and manufacturing services page covers how we structure sourcing and production for early-stage hardware products.
