Can I Make Money In CNC Machining?

• 1
  Where the money in a CNC job actually comes from
• 2
  Five cost drivers that decide if you make money in CNC machining
• 3
  Why job quantity and part mix change the answer
• 4
  Where outsourcing fits when you make money in CNC machining
• 5
  Which work actually pays: job profile compared

People ask can I make money in cnc machining as if the answer were a yes or no. It is a ratio. Every job is a small pile of billable time divided by a larger pile of time you never invoice: quoting, setup, probing, deburring, inspection, packing. A shop that charges USD 60 an hour for spindle time but spends 90 minutes setting up a 12-minute part is selling 12 minutes and giving away 90.

The margin lives in the gap between cycle time and total touch time. On a 3-axis mill running a simple aluminium bracket, cycle time might be 4 minutes and setup 25 minutes. Run 200 parts and setup amortises to 7 seconds per part. Run 5 parts and setup costs more than the cutting does. The same machine, the same drawing, two different businesses.

This is why quoting from a CAD model alone misleads people. The model tells you geometry. It does not tell you how many times a part gets picked up, flipped, re-clamped, measured and re-worked. On a mill-turn centre or a 5-axis machine, one setup can finish five faces, and that collapses touch time. On a 3-axis machine, the same part may need four fixtures and four operators' worth of attention.

So the first question is not how much a machine costs per hour. It is how many parts share one setup, and how much of the tolerance work can be done in-process instead of afterwards. Get that wrong and no hourly rate saves the job.

Tolerance is the first driver, and it is the one most often priced wrong. A part held to ±0.1 mm can come off the machine and go straight into a box. A part held to ±0.005 mm needs a temperature-stable room, sharp tooling, and an operator watching tool wear. The cutting time barely changes. The inspection time triples.

Material is the second. Aluminium 6061 machines fast and forgiving. Stainless 316 work-hardens if the feed is too light, so you cut deeper and slower and still change inserts more often. Titanium Ti-6Al-4V and Inconel add heat, tool wear and risk. The chip volume is the same; the cost per chip is not.

Setup and fixturing come third. Soft jaws and a vise are cheap. A custom fixture with a rotary table, a tailstock and a probe routine is not, and it is usually a one-off cost that only pays back over a run. The fourth driver is scrap. A near-net casting or forging that costs more than the machining time turns every mistake into a large loss. Bar stock is more forgiving.

Finishing is the fifth, and it is where quotes go quietly wrong. Anodising, electroless nickel, powder coating, bead blasting and laser marking all add handling, outside processing and lead time. Masking a hardcoat anodised surface for a tight bore is real labour. If the finish is not in the quote, the profit is not in the job.

One part and ten thousand parts are different industries. At one prototype, you are selling engineering judgement, speed and a DFM report, not spindle hours. At ten thousand parts, you are selling cycle time, tool life and material price. A shop that is good at one is often mediocre at the other, because the skills and the scheduling logic do not transfer.

Mix matters just as much. A shop running twenty different parts a month spends a large share of its week on programming, setup sheets, first-article inspection and paperwork. A shop running three parts on repeat can tune feeds, dedicate fixtures and hold the same cycle for months. The second shop usually shows better margins at the same hourly rate.

This is the practical reason many small shops chase repeat work in a narrow niche: medical housings, EV brackets, robot end-effectors, RF enclosures. Narrow work means fewer surprises, faster quoting and reusable tooling. Broad work means you learn fast and quote slowly. Neither is wrong, but they produce very different bank balances.

The boundary condition is simple. If setup time is under about 10 percent of the run time, you are in volume territory and should be optimising cycle time. If setup is over half the run time, you are in prototype territory and should be charging for engineering and speed, not for minutes at the spindle.

Not every shop owns every process. Most profitable ones buy out what they cannot keep busy: heat treatment, hardcoat anodising, precision grinding, EDM, injection moulding, die casting, sheet metal. The margin question then becomes whether you can manage those suppliers without losing control of tolerance and date.

A supplier that ships two days late destroys a schedule that was quoted at three to five days. So the working rule is to qualify the outside process the same way you qualify a machine: first-article report, capability data, and a written recovery plan. Cheap outside processing with weak documentation is expensive.

For engineers buying machined parts rather than making them, the same logic runs in reverse. You are not paying for spindle minutes. You are paying for a shop that has already solved the setup, the fixture, the tool path and the inspection method for a part like yours. That is what a quote from a capable supplier actually buys.

GreatLight runs 127 high-precision CNC machines across three plants, including 16 simultaneous 5-axis machining centres, 12 four-axis mills, 27 three-axis machines and 16 mill-turn centres, with a maximum processing size of 4,000 mm. Finishing, inspection and documentation sit under one roof in Dongguan and Singapore, which removes most of the hand-off risk described above.