
Sourcing machined parts from China used to be a straightforward cost decision: roughly half the Western price, a long lead time, and a quality risk you managed with a first article. That framing is out of date. Freight costs have moved, duty structures have changed, Chinese domestic wages have risen, and the supplier base has bifurcated into shops that can hold ±0.005 mm and shops that cannot.
What follows is an honest account of where the economics stand, what still goes wrong, and how to structure an order so the risks are visible before they are expensive.
The gap is real, but it is smaller and more specific than the folklore. On a moderately complex machined aluminum part at 50–500 pieces, landed cost from a competent Chinese supplier is typically 30–55% below Western European or North American pricing.
That gap comes from four things, and knowing which applies to your part tells you whether it will persist:
Labour cost — still materially lower, though the gap has narrowed as Chinese machining wages rose. This affects parts with high manual content: deburring, inspection, assembly, packaging.
Machine amortisation — a shop running three shifts on a machine bought for domestic work amortises it faster than a Western shop running one shift.
Material buying power — large Chinese shops buy aluminum and stainless in volumes that a small Western job shop cannot match, and China is a major producer of both.
Supply chain density — anodizers, heat treaters and platers are clustered, so subcontract steps are hours away rather than days.
Where the gap collapses: single-piece orders (freight and communication overhead dominate), very simple parts (local shops compete on convenience), and parts with extreme tolerance or exotic material (the qualified supplier pool shrinks and pricing rises accordingly).
Air freight rates have been volatile, and sea freight — while cheaper — now carries longer and less predictable transit. For a 200 kg shipment, the difference between a good and a bad freight quarter can exceed the part price difference on simple components. Build freight into the comparison at current rates, not rates remembered from a few years ago.
There are now clearly two tiers. Tier one: shops with 5-axis capability, CMM inspection, ISO 13485 or AS9100, English-speaking engineering, and prices that reflect it. Tier two: shops competing on price with 3-axis equipment, limited metrology, and variable communication. Both can be right for different parts — the mistake is expecting tier-two pricing with tier-one process control.
Chinese shops increasingly have domestic customers for EV, battery, and consumer electronics work. A small foreign order that would once have been welcomed may now be scheduled around larger domestic commitments. This shows up as longer lead times rather than higher prices, and it is worth asking about directly.
Origin rules, forced-labour screening and tariff classifications have become more demanding in several markets. For most machined components this is administrative, but it is no longer something to discover at customs. Confirm the HS code and the origin documentation with your forwarder before the first shipment.
| Risk | How it shows up | How to cap it |
|---|---|---|
| First batch conformance | Dimensions off, finish wrong | FAIR with measured values, your approval as a gate |
| Batch-to-batch drift | Part 1 fits, part 300 does not | Sampling plan in the PO, retained records, lot marking |
| Material substitution | Wrong grade or temper; part behaves differently | Mill certificate matched to lot number on delivered parts |
| Communication lag | Questions answered a day later, assumptions made | Named engineer contact, DFM notes returned with the quote |
| Schedule slip | Finishing queue, material procurement | Lead time broken down by stage; split shipments |
| Payment exposure | Deposit lost if supplier fails | Staged payments tied to inspection milestones; small first order |
The most effective single control is not a contract clause. It is ordering one representative part, inspecting it yourself, and judging four things: dimensional conformance, surface and edge quality, documentation quality, and how the supplier responds to a nonconformance. That costs a few hundred dollars and replaces most of the guesswork.
Start with 1–5 pieces. Not to save money — to see the process. Judge documentation as much as the part.
Send a complete RFQ. 3D model, 2D drawing with tolerances and datums, material spec with temper, finish spec, quantity breaks, inspection requirement.
Read the DFM response. A supplier who marks up your drawing has engineering depth. One who returns a price only has a quoting template.
Approve the first article in writing. Make production wait for it. This one clause prevents most batch-level failures.
Then scale. 50, then 500. Each step is a chance to catch drift before it is expensive.
China is not automatically correct. Local sourcing wins when:
The part is needed in days, not weeks — freight alone decides it
The design is still changing and you need same-day DFM conversations
The quantity is under about 10 pieces and the geometry is simple
The part is very large or very heavy, where freight scales badly
Regulatory or customer requirements mandate domestic manufacture
A useful discipline is to quote both and compare landed cost at your real volume, including freight, duty, and the internal cost of managing a distant supplier. That last item is real and usually ignored: a Chinese supplier needs clearer drawings, more explicit specifications and more documentation than a local shop you can visit.
Not inherently. The range is wider than in mature markets — the best Chinese shops match anyone, and the worst are worse than anything you would encounter locally. The variance is the risk, and it is managed with a first article and a sampling plan, not with assumptions.
For 50 machined aluminum parts with anodizing: 10–14 days production, 3–5 days air freight, plus customs clearance. Six to eight weeks from PO for a 500-piece order is a reasonable planning assumption.
Typical terms are 30% deposit and 70% before shipment. For a first order, negotiate a smaller deposit and tie the balance to your approval of the first article or to a pre-shipment inspection. Avoid 100% upfront on a first order.
If you buy frequently from multiple suppliers, a good agent earns their margin on inspection and coordination. If you buy one part occasionally, direct contact with a supplier who speaks your language is usually simpler and cheaper.
Use an NDA, share only what is needed for the quote, and avoid sending complete assemblies to a supplier who only makes one component. For genuinely sensitive designs, split the drawing so no single supplier sees the whole system.
Most sourcing comparisons fail at the arithmetic stage, because they compare unit price rather than landed cost. The components that get left out are consistently the same ones.
| Cost element | Typical magnitude | Often forgotten |
|---|---|---|
| Unit price × quantity | Baseline | — |
| Tooling and setup | One-off, $0–$1,500 for machined parts | Yes |
| Freight | 3–8% of order value by air; 1–3% by sea | Sometimes |
| Duty and taxes | 0–12% depending on classification and destination | Often |
| Inspection on arrival | 1–3% of order value, more for critical parts | Yes |
| Internal management time | 5–15 hours per order for a distant supplier | Almost always |
| Inventory carrying cost | Larger batches mean more stock on the shelf | Almost always |
| Quality escape cost | A rejected batch, a line stoppage, a field return | Until it happens |
Two of these deserve emphasis because they decide borderline cases.
Internal management time is real even though it appears on no invoice. A distant supplier needs clearer drawings, more explicit specifications, more documentation and more follow-up than a shop you can drive to. On a small order, that overhead can equal the price difference — which is why the economics of sourcing abroad improve with order size and with repeat business.
Inventory carrying cost follows from the quantity breaks. If the price at 1,000 is 30% below the price at 200, the saving is partly consumed by holding 800 units you did not need yet, at a carrying cost of roughly 20–25% a year including capital, space and obsolescence risk. The honest comparison is between the quantity you can actually consume and the price at that quantity.
Run both numbers before deciding. On a 200-piece order the gap may be 20%; on a 2,000-piece order with a stable design, it may be 50% — and the decision is different in each case.
The cost advantage is real but narrower than it was, and concentrated in parts with meaningful labour content at quantities above a few dozen. The risks are manageable and well understood — first article, sampling plan, material certificates, staged payments. What does not work is treating a distant supplier as if proximity made specification optional.
We work with overseas customers on exactly this basis: DFM notes with the quote, FAIR before production, material certificates matched to lots. See our capabilities, or send a drawing with your target volume and we will quote landed cost against your alternative.