Japan or the US: choosing where to build your first production run
A decision framework for choosing a first manufacturing region on a hardware program — cost, capability, lead time and the hybrid pattern that usually wins.
Run it as a decision matrix, not a preference
Founders often pick a manufacturing region by familiarity — the founder's home country, the factory a contact works at, or whichever conference they last attended. On a first production run the stakes are too high for that. Four dimensions decide it: unit cost at your real volume, capability for your hardest part, lead time from design change to shipped product, and risk concentration.
Score each candidate region against all four for your specific bill of materials. A region can lose on cost and still win overall because it owns the one capability your program cannot risk.
When Japan wins
Japan wins when tolerance and consistency are the product. Precision motors, battery components, tight-tolerance castings and tooling that has to hold spec across hundreds of thousands of cycles are its home ground, and the supplier depth there is unmatched.
The costs are longer first-article lead times, higher piece prices on commodity parts, and a slower relationship-building phase with suppliers. If your program's hardest problem is manufacturing precision, pay those costs.
Related capability: EV powertrain and battery pack engineering · Vehicle mechanical engineering services
When the US wins
The US wins on iteration speed, tariff and freight risk, domestic-content requirements, and low-to-mid volume mechanical work — CNC machining, harnesses, sheet metal, and final assembly close to your engineering team. When the design will still be changing during the first production run, physical proximity between engineering and the line is worth real schedule.
Where it loses is cells, deep electronics ecosystems, and anything where piece price at volume decides the business case.
Related capability: Vehicle industrialization and production ramp · Vehicle electronics and software development
The hybrid that usually wins
Most programs we advise end up split: the highest-precision subsystems sourced from Japan, mechanical components and final assembly in the US, and the choice revisited at each volume step. The split adds coordination overhead — two quality systems, two sets of supplier relationships — so it only works when the interfaces between subsystems are cleanly defined.
If your architecture cannot support that split, that is itself useful information: it usually means subsystem boundaries were drawn by convenience rather than by engineering, and redrawing them pays off regardless of where you build.
Related capability: Industrial design for electric vehicles · Vehicle validation and homologation
Working on a program like this?
We help OEMs and venture-stage teams take EV programs from sketch to production.
Talk to LAND Labs