CNC processing costs per unit guide

• 1
  Key takeaways
• 2
  What moves CNC processing costs per unit
• 3
  How CNC processing costs per unit are actually built
• 4
  Quantity: the single biggest lever on unit cost
• 5
  Geometry, tolerance and setup count
• 6
  Material choice and how much of it you remove
• 7
  Finishing, inspection and the costs outside machining
• 8
  How to compare quotes without getting fooled

Every unit price is a fixed part and a variable part. Fixed costs are programming, fixture design, first-article inspection and machine setup. They are charged once per order, so dividing them by quantity is what makes a prototype expensive and a 1,000-part run cheap. If a prototype carries a large share of setup, the same part at 1,000 pieces can fall to a small fraction of that number.

The variable part is cycle time multiplied by the machine rate, plus material, tooling wear and finishing. Cycle time is where engineering decisions bite. A part with deep pockets, thin walls or organic 3D surfaces runs the tool along far more path length than a plate with drilled holes.

Material is usually quoted at the billet weight you actually buy, not the finished part weight. A bracket machined from a 1 kg block down to 150 g pays for the whole block. Titanium and nickel alloys cost more per kilogram and also cut slower, so they hit the price twice.

Labor, maintenance, software licensing and power sit in the hourly rate. A shop running simultaneous 5-axis centers and automated tool changers spreads those costs over more spindle hours, which is why two shops with similar machines can quote differently.

• 1
  FixedProgramming, fixturing, setup, first article
• 2
  VariableCycle time, material, tool wear, finishing
• 3
  HiddenInspection time, packing, freight, payment terms

Setup is not free and it does not shrink with volume. Whether you order 1 piece or 500, the machine still needs to be fixtured, the program proven and the first part measured. That is why a one-off prototype can cost many times the unit price of the same geometry at volume.

The break-even curve is steep at first and then flattens. Going from 1 to 10 parts changes the unit price a lot. Going from 500 to 1,000 changes it much less, because material and cycle time now dominate. If a quote barely moves between two quantities, the fixed costs were already small.

Buyers often split a program into small releases to reduce inventory. That is a real benefit, but each release carries its own setup. Consolidating two orders into one run, or ordering a year of spares at once, is usually the cheapest change you can make without touching the design.

We run from one prototype to 10,000+ part runs with no minimum order quantity, so the same process and inspection standard applies at both ends of that range.

A simple 2D profile with through holes cuts fast. Pockets, undercuts, deep ribs and thin walls force smaller tools, lighter passes and more retracts. Thin walls also vibrate, so the programmer slows the feed to hold the dimension. Both effects show up as cycle time.

Tolerance is a cost multiplier, not a checkbox. General dimensions at ±0.1 mm run at normal feeds. A critical bore at ±0.005 mm needs a finishing pass, a temperature-stable measurement and often a dedicated gauge. Tighten only the features that mate with something else.

Feature count matters more than part size. One 400 mm housing with twelve features can cost less than a 60 mm manifold with forty intersecting ports, because each feature adds tool changes and verification.

Setup count is the quiet killer. A part machined in one 5-axis setup avoids the re-fixturing error and the labor of three separate 3-axis operations. Moving work from three setups to one often cuts both cost and scrap.

Material cost is billet weight times price per kilogram, plus the cutting penalty. Aluminum 6061 machines fast and is cheap, which makes it the default for housings and brackets. Stainless 304 and 316 are slower and work-harden, so tool life and cycle time both rise. Titanium TC4 and Inconel sit at the top: expensive stock, low cutting speeds and short tool life.

Removal volume is the number most drawings hide. A part that starts as a near-net forging or a standard plate thickness removes less material, uses fewer passes and generates less chip. If a design removes 80% of the block, ask whether a different stock form or a casting would remove less.

Wall thickness drives both cost and risk. Very thin sections need support, gentle feeds and sometimes a stress-relief step between roughing and finishing. If the wall is not functional, thickening it by 0.5 mm can remove a whole operation.

Standard stock sizes matter. Choosing a billet that matches an available plate or bar size avoids a cutting charge and reduces scrap, even when the material grade stays the same.

Machining is often only part of the invoice. Anodizing, plating, powder coating, bead blasting and laser marking are separate operations with their own setup and minimum charges. A color anodize batch has a minimum load, so a handful of parts can cost more per piece than the machining.

Surface finish is specified per feature, not per part. A sealing face at Ra 0.2–0.8 μm needs polishing or fine finishing. An internal bracket at Ra 1.6–3.2 μm comes straight off the machine. Writing one finish note for the whole drawing usually over-specifies the easy features.

Inspection scales with risk. A general tolerance part may need a first-article check and a few dimensional spot checks. A medical or aerospace part at ±0.005 mm may need 100% inspection with reports and traceability. Both are legitimate, but they are different products and should be quoted as such.

Certification requirements also add administrative cost. ISO 9001:2015, IATF 16949:2016, ISO 13485:2016 and ISO 27001:2022 each bring documentation and process controls that a general machine shop does not carry.