How to Open a CNC Machine Shop

• 1
  Key takeaways
• 2
  What a CNC machine shop actually sells
• 3
  Choosing a niche for a new CNC machine shop
• 4
  Machine mix, floor plan and services you need to quote
• 5
  Tolerance, materials and inspection planning
• 6
  People, quality system and the paperwork buyers ask for
• 7
  Finding the first customers and keeping them
• 8
  Scaling a CNC machine shop without breaking quality
• 9
  How to open a CNC machine shop: 7 steps in order
• 10
  Which machine should a new shop buy first?

Before you sign a lease, be clear about the product. A CNC machine shop does not sell machining time. It sells parts that fit the assembly, arrive on the promised day, and come with paperwork the buyer can file. That last part is what separates a shop that survives year two from one that closes.

So the first question is not which machine to buy. It is which part family you can repeat. A shop that runs the same aluminium housing every month learns the fixture, the tool wear, and the inspection routine. A shop that runs a new shape every week never gets faster.

Write down the parts you expect to make: size, material, tolerance, quantity per year. Then pick the machine that covers 80 percent of that list. The remaining 20 percent can be subcontracted until volume justifies another spindle.

• 1
  Part familyOne or two shapes you can quote without re-quoting every time.
• 2
  Tolerance bandGeneral machining at ±0.05 mm costs far less to hold than ±0.005 mm.
• 3
  Order patternPrototype runs of 1–50 parts behave very differently from 10,000-piece runs.

General machine shops compete on price. Specialists compete on fit. Pick a niche where you can answer a technical question the buyer cannot answer alone. Automotive brackets, medical instrument frames, robot end-effectors, and heat-sink housings are common starting points because the drawings are stable and the quantities repeat.

Talk to ten local product engineers before you buy anything. Ask what they wait on. Long lead times, poor surface finish, missing inspection reports, and suppliers who disappear after the first order come up again and again. Those gaps are your market entry.

Do not pick a niche purely because the margin looks high. A medical implant shop needs cleanroom discipline and traceability from day one. If you cannot fund that, start with industrial machinery parts and move up as your quality system matures.

• 1
  Aerospace and defenceTight tolerances and full traceability. High barrier, high retention.
• 2
  Automotive and EVVolume and IATF 16949 expectations. Good for repeat production.
• 3
  Industrial machineryWider tolerances, shorter approval cycles. A practical first niche.

Most new shops buy a 3-axis mill first. That is fine for plates and simple pockets. It is not enough for parts with features on five faces, because every extra setup adds stack-up error and labour. If your niche includes impellers, medical housings, or engine components, plan for a 5-axis centre or a 4-axis mill with a rotary table.

Travel matters more than spindle speed at this stage. A 500 × 500 × 450 mm envelope handles a large share of brackets and covers. A 750 × 1,150 × 550 mm machine opens up bigger frames but needs floor space, power, and a crane. Measure the door before you buy the machine.

Plan the floor around material flow, not around the electrical panel. Raw bar stock in, saw, first op, second op, deburr, inspection, packing. If a part crosses the shop three times, you lose margin on every order. A 200 m² floor can run four machines well if the layout is clean.

Decide early which services you quote yourself and which you broker. Turning, milling, surface finishing, and inspection are core. Anodising, plating, and heat treatment are usually better outsourced until volume justifies the tanks and permits.

• 1
  3-axisPlates, pockets, simple housings. Lowest entry cost.
• 2
  4-axis with rotary tableCylindrical parts and multi-face work in one setup.
• 3
  5-axis simultaneousComplex contours, undercuts, and tight position tolerance.
• 4
  Mill-turnShaft-type parts that need turning and milling without re-fixturing.

Set your tolerance promise from the process capability, not from the sales sheet. A well-maintained machining centre holds ±0.005 mm on a stable part with the right fixture. It will not hold that on a thin-wall aluminium cover without chatter control and a stress-relief step.

Surface finish follows the same rule. As-machined surfaces land around Ra 1.6–3.2 μm. A high-quality finish runs Ra 0.8–1.6 μm and usually needs a finer step-over or a finishing pass. Below Ra 0.8 μm you are in polishing or lapping territory, and the cost curve turns steep.

Inspection is where new shops cut corners and lose customers. Plan for raw material verification, in-process checks, and a final inspection before shipment. A basic toolkit covers calipers, micrometers, height gauges, pin gauges, and a surface roughness tester. A CMM is the next step when position tolerance enters the drawing.

Materials set the cutting parameters and the paperwork. Aluminium 6061 and 7075 machine fast and are forgiving. Stainless 316L and 17-4PH work-harden, so keep the feed per tooth up and never dwell. Titanium Ti-6Al-4V and Inconel need rigid setups, lower surface speed, and plenty of coolant.

• 1
  Aluminium6061, 6061-T6, 7075, 6082. Good first material for a new shop.
• 2
  Stainless303, 304, 316L, 17-4PH. Watch work hardening on light passes.
• 3
  Steel1018, 1045, 4140. Predictable, but slower and heavier on tools.
• 4
  PlasticsPOM, PEEK, PC. Clamp lightly. They move under pressure.

You can start with two or three people. One programmer who also sets up, one operator, and one inspector who is not the operator. That last separation matters. Self-inspection is normal in small shops, but final inspection should be done by someone who did not cut the part.

Training is not a one-time event. Tool wear, fixture maintenance, and drawing revision control need a routine. A short weekly review of scrap, rework, and late jobs catches problems before a customer does.

A quality management system does not have to be heavy at the start. Write down how you control drawings, how you approve a first article, how you handle a non-conformance, and how you keep records. ISO 9001:2015 is the usual first certificate because most buyers accept it. Automotive work adds IATF 16949:2016. Medical work adds ISO 13485:2016. Data handling under ISO 27001:2022 matters when you hold customer CAD files.

Buy insurance before the first chip. Liability cover for defective parts, property cover for the machines, and workers compensation. One scrapped batch on a critical part can cost more than the premium.

• 1
  First article inspectionFull dimensional report on the first part of every new revision.
• 2
  Revision controlEvery drawing gets a number, a date, and an owner.
• 3
  Non-conformance logWhat went wrong, what you did, what changed. Buyers ask for it.

New shops chase big OEMs too early. Those accounts take months to approve and often demand a quality history you do not have yet. Start with product design studios, machine builders, and robotics integrators who need 5 to 200 parts and cannot wait six weeks.

Your first quote should include a short manufacturability note. Flag the feature that will be expensive, suggest a tolerance relaxation if the function allows it, and state the lead time honestly. Engineers remember the supplier who told them the truth about a bad design.

Keep the pipeline moving. Ten quotes a week is a healthier habit than one large bid. Track win rate, average order value, and the reason you lost. If you lose on price every time, your niche is wrong or your process is slow.

Retention is a process, not a feeling. Send the inspection report. Confirm the delivery date two days before shipment. Ask what the next revision looks like. Shops that do these three things rarely lose an account to a cheaper competitor.

• 1
  Quote with DFM notesPoint out the costly feature and offer an alternative.
• 2
  Deliver early on the first orderIt buys trust that a discount cannot.
• 3
  Track the loss reasonPrice, lead time, or capability. Each needs a different fix.

Scale after the process is repeatable, not before. The signal to add a machine is a queue you cannot clear with overtime, not a trade show discount. Adding capacity to an unstable process multiplies the rework instead of the revenue.

Add the second shift before the fifth machine. A second shift spreads the fixed cost of the building, the compressor, and the quality system across more hours. It also exposes weak documentation, because the night crew cannot ask the programmer in person.

Outsource the finishing steps first. Anodising, plating, heat treatment, and laser marking are easy to broker and hard to run. Bring them in-house only when the volume is steady enough to justify the permits and the operator.

Keep the quality system ahead of the customer list. If you want automotive work, build toward IATF 16949:2016 before the audit, not after. If you hold customer CAD files, ISO 27001:2022 practices protect both sides.

• 1
  Add shifts before machinesSame capital, more output, lower fixed cost per part.
• 2
  Broker finishing firstLow capital, high flexibility while volume is uncertain.
• 3
  Grow the certificate with the customerISO 9001 first, then IATF 16949 or ISO 13485 as the market demands.