Can You Make Money With a CNC Machine?

A used 3-axis mill might list at $15,000–$40,000, and tooling plus CAM software adds another $5,000–$20,000. Those numbers are easy to find. The costs that sink small shops are harder to see.

Power draw on a mid-size VMC runs roughly 10–50 kWh. Material waste on a first-run job can hit 15% scrap before the process is dialed in. Tool wear on stainless or titanium is not a rounding error either—a single 6 mm carbide end mill may last a few hours in 316L and days in 6061. Add spindle rebuild intervals, coolant disposal, and compressed air, and the hourly rate you charge has to cover far more than the mortgage on the machine.

The clean way to test it: add machine payment, tooling, power, coolant, floor space, and your own time into one hourly number. Then compare that number against what the market pays for the parts you intend to cut. If the gap is thin, the business model is wrong, not the machine.

Labor is the line most often left out. Even a one-person shop spends unpaid hours on quoting, deburring, packing, and chasing material. Those hours come out of the same margin.

• 1
  Machine paymentSpread over realistic billable hours, not 24/7 fantasy.
• 2
  Tooling and CAMConsumable, not a one-time purchase.
• 3
  Scrap and reworkFirst-run scrap on tight-tolerance parts is common.
• 4
  Unpaid adminQuoting and packing hours still cost money.

The parts that pay well share a few traits: tight tolerances, awkward geometry, small batch sizes, and a customer who needs them fast. A bracket that any job shop can run in 6061 at ±0.1 mm is a commodity. A 5-axis impeller with ±0.005 mm features and a two-week deadline is not.

Low-volume prototype runs are the sweet spot for a small shop. Setup is a fixed cost, so the margin per part rises as batch size drops and complexity rises. One-off fixtures, medical instrument housings, and motorsport brackets all fit this pattern.

Commodity work still makes money, but only at volume and with automation. A 10,000-part run of a simple aluminum spacer needs bar feeders, lights-out operation, and tight cycle times. A manual shop competing on that job is bidding against machines that run unattended overnight.

Parts with secondary operations are a trap worth naming. A part that needs anodizing, heat treat, and laser marking has three outside vendors in the chain, each adding lead time and a markup you may not be able to pass through. Either quote those operations honestly or pass on the job.

• 1
  Good fitPrototypes, tight tolerances, complex geometry, short runs.
• 2
  Marginal fitSimple parts at mid volume with no fixturing advantage.
• 3
  Poor fitCommodity parts at high volume with no automation.
• 4
  Watch outParts needing three or more outside processes.

Start from cycle time, not from what the last job paid. Break the part into setup, roughing, finishing, and deburring. Estimate each one, then add a scrap allowance. If the estimate is within 20% of a real quote, your process model is probably good enough to bid on.

The hourly rate should cover machine, labor, overhead, and profit. A shop charging $75/hour on a $30,000 machine needs roughly 400 billable hours just to cover the machine, before overhead. That is the math that matters, and it is why an idle spindle is so expensive.

Quote the setup as a separate line on prototype jobs. Customers accept a $400 setup fee on a $200 part far more easily than a $600 part price with no explanation. Transparency on setup also protects you when the customer changes the drawing after the first article.

Revise quotes when the drawing changes. A tolerance change from ±0.05 mm to ±0.005 mm can double the cycle time on a finishing pass. Free revisions are a margin leak that shows up months later.

• 1
  Cycle time firstSetup, roughing, finishing, deburring, scrap.
• 2
  Rate covers everythingMachine, labor, overhead, profit.
• 3
  Show setup separatelyCustomers accept it more easily than a padded part price.
• 4
  Re-quote on revisionTighter tolerance means more passes, more time.

The first loss is idle time. A machine that runs 20 hours a week cannot carry the same overhead as one that runs 60. Most new shops underestimate how much time goes to sales, quoting, and material sourcing before the spindle turns.

The second loss is quoting optimism. Cycle times estimated from a CAM simulation ignore tool changes, chip evacuation pauses, and the operator checking a feature mid-cut. Add 15–25% to simulated cycle time on a new part and you will be closer to reality.

The third loss is scope creep. A customer asks for a small change, then another, and the original quote no longer applies. Written change orders are not bureaucracy—they are how you keep the job profitable.

The fourth loss is tooling choice. Running a cheap end mill to save $20 often costs three times that in scrapped parts and extra finishing passes. On stainless and titanium, tooling is not the place to economize.

• 1
  Idle spindleOverhead runs whether the machine cuts or not.
• 2
  Optimistic cycle timesAdd 15–25% over the CAM simulation.
• 3
  Scope creepChange orders keep the job profitable.
• 4
  False tooling savingsCheap cutters cost more in scrap and rework.

Buying a CNC machine only makes sense when you have enough repeat work to keep it busy. A shop with steady orders for the same family of parts can justify the capital. A team that needs 20 prototype parts twice a year usually cannot.

Outsourcing lets you test a market before committing capital. You quote the job to a partner, learn the real cycle time and cost, and decide whether the demand is there. If it is, you buy later with better data.

There is also a capability argument. A 3-axis machine cannot reach the features on a 5-axis part, and a lathe cannot mill a prismatic housing. Partnering with a shop that has the right machine for the job is often cheaper than buying one you use twice a month.

GreatLight runs 127 high-precision CNC machines across three plants, including 16 simultaneous 5-axis centers and a 4,000 mm maximum processing size. We quote and provide free DFM analysis within 12 hours, and production can start within 24 hours. No minimum order quantity, from one prototype to 10,000+ part runs.

• 1
  Buy whenSteady repeat work keeps the spindle busy.
• 2
  Outsource whenDemand is unproven or volumes are low.
• 3
  Capability gap5-axis and turn-mill work often goes outside.
• 4
  Test firstQuote the job, learn the real cost, then decide.