The short answer
The logic sounds airtight: if mass production will happen in China, why not start there and save money? Sometimes that is right — but for early prototypes it is usually premature optimization, because the currency of the prototype phase is learning speed, not unit price. The useful rule: build early iterations close to your team, and use overseas suppliers once the design is stable enough that you are buying quantity, finish, or a process you cannot get locally.
Why first models belong close to you
Iteration lives on cycle time: build, test, fix, rebuild. When each loop crosses an ocean, it absorbs international shipping, customs, a twelve-hour time difference, and a language gap. A two-day loop becomes a two-week loop, and three loops cost you a month and a half at the stage where time is your scarcest asset.
There is a second, less obvious cost. Early models change constantly and are half-specified by nature — exactly the situation where standing next to the machinist or printer is worth more than the price difference. A local shop reads intent from a sketch; a remote shop can only read the drawing you sent, and will build precisely what it says. For the local toolkit, see rapid prototyping techniques and CNC machining for prototypes.
When overseas prototyping genuinely wins
- Pre-production and pilot runs. Once the design is frozen and you need dozens to hundreds of well-finished units for beta testing, trade shows, or investor demos, the cost advantage is large and the lead time is tolerable.
- First-article samples off tooling. The initial shots from a new mold are themselves a kind of prototype, and they can only come from wherever the tool lives. This is the natural handoff described in going from prototype to production.
- Custom subassemblies and components. Displays, battery packs, small motors, custom cables, silicone parts, and printed packaging are fast and inexpensive to get in sample quantities from Asian suppliers.
- Processes with no accessible local option. Certain finishes, low-volume die casting, or specialized rubber and glass work can be genuinely hard to source domestically at prototype quantities.
- Trialing your future manufacturer. Having your production candidate build a sample run is the cheapest audition you will ever get, and the process knowledge stays in the supply chain.
The real cost math
Overseas prototype quotes often land at a fraction of US shop pricing, and that number is real — but it is not landed cost. Add:
- Express freight for a small parcel, which frequently costs more than the parts themselves.
- Duties and fees. Samples are not automatically duty-free; classification, tariffs, and broker fees apply — see import duties and tariffs from China.
- Iteration overhead. One extra loop caused by a misread drawing usually erases the entire saving.
- Your own time spent specifying, chasing, and re-explaining across time zones.
Typical honest ranges: a handful of machined or printed prototype parts from an overseas shop can come in at a small fraction of domestic pricing, arriving in roughly two to four weeks door to door; the equivalent local job costs meaningfully more but can be in your hands in days. For a stable design in pilot quantities, the overseas advantage compounds; for a design still moving, it evaporates. If cost is the binding constraint, work through how to cut prototype costs before changing continents.
Realistic lead times
- Quotation and DFM feedback: a few days, longer around national holidays.
- 3D printed or machined parts: roughly one to two weeks in production.
- Soft or bridge tooling with sample shots: several weeks.
- Express air freight and customs clearance: several days to about a week.
- Plan for a full stop of two to four weeks around the Lunar New Year period — the shutdown is unavoidable and the surrounding weeks are degraded.
The rules if you do it
- Send a complete, unambiguous package. Dimensioned drawings with tolerances, critical dimensions flagged, material and finish specified by standard designation, and no "same as before" references. A remote build lives or dies on definition quality.
- Protect yourself before files leave. Confidentiality, non-use, and non-circumvention terms first — see NNN agreements for China manufacturing. For the earliest and most sensitive builds, consider sending only selected parts rather than the complete assembly.
- Validate twice. Ask for photos and video of the parts, and of the unit functioning, before shipment. Then run your own incoming inspection with a written acceptance list.
- Prefer the supplier with a future. Build samples at the shop that is a real candidate for production, so the learning is not thrown away.
- Ship in one consolidated batch, and order spares of anything fragile. A second shipment for one missing bracket costs a week.
- Keep firmware and software in-house. Send hardware to be built; do not hand over your source.
What actually goes wrong
The recurring failures are not exotic. Silent material substitution to a cheaper resin. A tolerance quietly opened up because it was inconvenient. A part built exactly to a drawing that contained an error nobody flagged. Cosmetic standards interpreted generously. Almost all of it traces back to specification gaps rather than bad faith, which is why the drawing package matters more than the audit for prototype work. For hardware entering a formal build sequence, align your sample runs with the standard staging in EVT, DVT, and PVT explained.
The hybrid path that works
The pattern we use with clients: fast local iterations until the mechanism and ergonomics are proven, a formal design freeze, then a pilot batch overseas at the supplier who will produce at volume, followed by tooling. You pay for speed where speed is worth the most, and for unit cost where quantity makes it matter. More on staging is in our prototyping section.
Get the package right before you send it anywhere
Projects House builds prototypes for US clients both locally and through overseas partners, and prepares the drawing and specification package that makes a remote build come back correct the first time. If you are deciding where to build your next iteration, tell us about the project through our contact form.