CNC Machining Prices November 2, 2026

Read that table once and then set it aside, because on a typical machined part material is 12 to 22% of the piece price. A 1.86% move on 6061-T6 shifts a $25.10 part by about two cents. Everything below moves it by dollars.

A revision is not priced by how big it is. It is priced by how much of the shop's work it invalidates. Here is the ladder, worked on Part 1472 at 300 pieces.

The failure mode is predictable and it is a buyer behaviour, not a shop behaviour. Someone notices the fit problem on day 4, decides not to "disrupt the schedule", and sends the revision on day 21 - by which point the shop is past halfway and the same piece of information costs $4,144 instead of $260. Telling the shop early never makes a programme later; it is the only thing that makes it cheaper.

The Two Clauses That Close This

Require a first-article gate. Specify that production does not start until you sign off the FAI report and dimensioned photos. The gate costs $240 on Part 1472. It is what lets you catch a fit problem at stage 4 ($2,060) instead of stage 7 ($7,530). $240 to avoid up to $5,470.

Authorise the shop to stop. Write into the PO that on receiving any revision the shop will notify you within 24 hours with a cost and schedule number, and will hold work until you respond in writing. Without this clause the shop has two bad options - guess, or keep running into the scrap - and it will pick the one that keeps its schedule, which is not the one that protects your budget.

Freeze the print for the lot. Any change that is not a safety or fit-blocking issue goes on the next lot. On a 600-per-year part released twice, the wait is 26 weeks on paper and about 8 weeks in practice.

Machine shops quote machining accurately. Finishing is where the estimate goes soft, partly because much of it is subcontracted and partly because it is quoted as a single word - "anodize", "passivate", "plate" - when it is really a list of separate charges. Below is what each one actually costs, and the dimensional change it imposes on the part.

Two structural consequences follow, and both are easy to miss at quote stage.

Lot minimums invert the economics on small quantities. On a 20-piece order, a $145 anodize minimum is $7.25 per piece before the tank is even switched on. At 300 pieces it is $0.48. Finishing, not machining, is what makes small lots expensive - and it is why consolidating two small releases into one is almost always the cheapest move available to a buyer.

Every coating changes the part's dimensions. Hard anodize grows a surface by 5 to 13 microns, powder coat by 80 to 150 microns. If the print calls out ±0.05 mm on a shaft that gets hard anodized, the parts will not assemble - and the shop will correctly say the machined dimension was in tolerance. Name the coating on the print and state whether the tolerance applies before or after finishing. This single line prevents a class of dispute that is otherwise unwinnable after the fact.

The most common expedite request is "run my parts overtime". It is almost always the wrong lever, because overtime compresses the three days that are actually cutting metal and does nothing about the thirty-one days that are not.

Add it up and machining is 9% of the calendar. The other 91% is waiting - waiting for material, waiting for a machine, waiting for a vendor's tank. So the two levers that actually work are not the ones buyers reach for:

Reduce the number of external operations. Every outsourced step adds its own queue, and an outside vendor's queue averages 5 to 9 days regardless of how urgent your job is - you are not their only customer. Part 1472 has two external steps. Removing the laser-mark subcontract by moving the mark in-house removes one full queue: about 6 days for a $75 setup.

Ask for a partial delivery instead of a rush. Fifty pieces on day 12 and the balance on day 26 costs one extra setup, $340, plus split freight. A full-lot expedite costs +22 to 48% of the piece price - on 300 pieces at $25.10 that is $1,660 to $3,615. The partial delivery is 5 to 10 times cheaper and usually gets you parts sooner.

Specify stock sizes where the design allows. Non-stock material is 4 days of the 34 and costs $0.40–1.20/kg more. On a part where one dimension is not critical, letting the shop use a stock bar size removes those four days at zero cost.

Default shop terms read roughly: nonconforming product will be repaired or replaced at our option. Those last three words mean the party that made the defect chooses the remedy. On a lot where you need parts next week and the shop wants to rework them over three weeks, that is the entire dispute.

Two of those deserve expansion, because they are pure money that buyers leave on the table.

Rework Versus Replace: A Rule, Not A Debate

Rework is worth authorising when the rework cost is under 60% of replacement cost and the affected feature is not a sealing face, a fatigue-loaded radius, or a thread. Above 60%, or on any of those three feature types, replace. The 60% figure is not about money - it is about the fact that a reworked part has been handled, re-fixtured and re-inspected, and every one of those steps is a new opportunity for the defect you already found to come back in a different form.

Swarf Credit: The Line Nobody Writes

When a part is scrapped, the shop keeps the chips and the buyer has paid for metal that is now someone else's asset. On aluminium that is noise. On the expensive alloys it is not:

On a job consuming 400 kg of C11000 a year, that is $4,400 to $4,840. Write one line: "Swarf remains buyer's property; credit to be applied at [rate] per kg, or scrap returned at buyer's request." Shops accept it routinely because they were already selling it - the line simply decides who gets paid.

Every buyer knows they should have a second source. Almost nobody has one, because qualifying one is expensive and slow, and the cost is invisible until you need it. Here it is, made visible.

The fixture is the hostage. If the PO does not say the tooling is yours, moving the part means paying to rebuild it - and once the fixture exists, the shop has every reason to argue it was their investment. Add the clause before the fixture is built; after that it is a negotiation you will lose. It costs nothing to add.

Two cheaper ways to get the same protection:

Qualify on a part you can afford to lose. Give the second shop a simpler, non-critical part first. You learn their programming, their inspection discipline and their communication for the price of one fixture on a part that does not matter - and the critical part is then qualified against a supplier you already know.

Split a lot instead of running a qualification. Give the second source 30 pieces of the next release. You pay one incremental setup, about $340, and no PPAP. You now have 30 parts from a live second source and a real comparison of the two shops on the same drawing.

And the timing point that matters most: the 8 to 16 weeks is the reason to do this when you do not need it. Qualifying after a failure means eight to sixteen weeks of line-down.

1. Price validity, in writing. Thirty days on repeat orders, ten days minimum on new ones. Machine shops are holding five days or less right now in the current input market.

2. Material escalation, bounded. If the named index moves more than 5% from the base date, the material leg only adjusts. Never accept unqualified "price at time of shipment" - it transfers 100% of the metal risk and it is not a price, it is an option written against you.

3. Tooling ownership and release. Fixtures paid for by buyer remain buyer's property and are released on request.

4. Revision procedure. Shop notifies within 24 hours with cost and schedule impact; buyer approves in writing before work restarts.

5. First article is a deliverable. Report plus dimensioned photos; production does not start until written sign-off.

6. Nonconformance remedy. Buyer elects repair, replacement or credit; sorting and containment at supplier's cost; containment in 48 hours.

7. Swarf. Buyer's property with a stated credit rate, or returned on request, for copper, titanium and high-nickel alloys.

8. Packaging and in-transit risk. Specify separators, VCI where relevant, moisture indicator and crate standard, and state who owns damage in transit - that is an Incoterms question, not a packaging question.

The copper and brass softness is late, not new. C36000 at $8.55/kg reflects a decline that happened weeks ago; brass shop quotes carry a two to four week lag, so November and early December brass pricing will keep easing even if copper turns back up. Brass is the one material worth not locking this month.

Aluminium is the opposite. Three consecutive gains on 6061-T6 with inventory drawing down. If you have aluminium work to place this quarter, the metal leg is moving against you - place it rather than wait.

Queue time is a demand signal, and it is the honest one. Machine rates are not moving; queue length is. Ask what the current queue is before you ask for a discount. A shop quoting a two-week queue is hungrier than one quoting eight - and a hungry shop is where discounts live.

Titanium's move is two rounds old and a third is signalled. Gr5 bar at $35.10/kg is up 3.75% in a week. On titanium parts, push for a material-linked price with a ceiling rather than a fixed number, because no shop will hold a fixed price through a six-week queue on a metal moving that fast.

Finishing vendor capacity is the hidden constraint. If your anodizer's quote suddenly includes a rush premium without you asking for one, their tank is full - and that, not machine time, will set your real delivery date.

Request A Machining Quote