How Easy Is It to Start a CNC Machine Shop?

• 1
  Key takeaways
• 2
  What it actually takes to start a CNC machine shop
• 3
  Floor space, power and the hidden fixed costs
• 4
  Staffing and the first jobs you should turn down
• 5
  Quality records and finding the first customers
• 6
  The failure modes that close new shops
• 7
  Step by step: opening a CNC machine shop
• 8
  Which machine and process fits your first year?

The mechanics are not the barrier. A used 3-axis vertical mill, a CNC lathe, a compressor, a phase converter and a small granite surface plate can be installed in a two-bay unit within a month. Most people who ask how easy it is to start a CNC machine shop already know how to run the machines. The difficulty sits in the two years after the first chip.

Start with the part, not the machine. Take the five prints you expect to quote most often and write down three numbers for each: largest envelope, tightest tolerance, and the material. A bracket at 200 × 150 × 40 mm in 6061 with a ±0.05 mm callout needs a very different machine than a 316L manifold at ±0.01 mm. Buying a 4,000 mm travel machine for 200 mm parts adds floor space, power draw and depreciation you cannot quote back to the customer.

Tolerance drives cost faster than size. Moving a feature from ±0.05 mm to ±0.005 mm typically adds a finishing pass, a temperature-stable room, and an inspection step. That is three cost centers, not one. Decide early which tolerance band you will sell. A shop that quotes everything at ±0.01 mm will lose money on simple work and still miss the tight work.

Then look at spindle hours. A single-shift 3-axis mill running 60 percent utilization gives roughly 1,200 spindle hours per year. If your target revenue per spindle hour is 40 USD, one machine supports about 48,000 USD of work. That number, not the machine price, tells you how many spindles you need to cover rent and wages.

A 7,600 m² plant is not where anyone starts, but the cost structure scales the same way. For a small shop, plan around 80 to 120 m² for two machines plus a bench, a saw and material racks. Leave 900 mm of clearance around each machine for service access. Add a separate area for deburring and inspection; chips and metrology do not belong in the same corner.

Power is the line item people forget. A small VMC with a 10 to 15 kW spindle needs a 30 to 40 A three-phase feed, plus a compressor at 7 to 8 bar. If the building only has single-phase service, a rotary phase converter works for light duty but limits you to one machine at a time. Budget for a proper panel before you sign the lease.

Consumables and tooling run 3 to 8 percent of revenue in a healthy shop. A new shop usually spends more in the first six months because every holder, insert and drill is a first purchase. Keep a tool crib list from day one so you can see what disappears.

Insurance, coolant disposal and calibration are recurring. Calibration of a single micrometer and caliper set on a 12-month cycle is not expensive, but if you quote aerospace work you will need traceable calibration on everything that touches a critical dimension.

One experienced setup machinist can run two to three machines if the work is repeatable. Prototype work is different. A shop doing one-off parts needs a programmer who can also set up, because handing a new program to an operator every morning wastes the first hour of every shift. Budget one programmer-setup person per two prototype machines.

Hiring is slow. Machinists who can read a print, pick a datum, and prove a first article are not common, and they usually have options. A new shop competes on stability of hours and on the quality of its equipment, not on hourly rate alone. If you cannot offer consistent overtime, say so in the interview.

Now the jobs to refuse. Turn down any part where the customer cannot state the function of the critical dimension. You will chase a tolerance that does not matter and lose the margin on the features that do. Turn down one-off work in hard materials like Inconel or Ti-6Al-4V unless the customer accepts tooling cost as a separate line. Those jobs eat inserts and time.

Also refuse work that requires a process you do not control. If a part needs vacuum heat treat and you have no qualified supplier, you own the delivery date anyway. Subcontract it only after you have run one job through the supplier and seen their paperwork.

A new shop wins its first orders on one of three things: speed, tolerance, or material capability. Pick one and be honest about the other two. If you have a 3-axis mill and a lathe, do not advertise 5-axis work. Buyers check.

Inspection is the cheapest marketing you can buy. A first article inspection report with every dimension recorded, plus material certificates, turns a one-off job into a repeat order. Raw material check, in-process monitoring and final inspection before shipment are standard practice at any shop that wants to keep industrial customers.

Where do the first customers come from? Local machine builders, automation integrators, and product companies with a prototype backlog. These buyers usually need 5 to 50 parts fast, and they value a shop that answers email the same day. Keep a simple capability sheet: machine list, envelope, tolerance band, materials and finishing partners.

Expect a six to twelve month ramp. Month one to three is usually one customer and one part family. Month four to eight adds a second buyer. If nothing repeats by month twelve, the problem is usually positioning, not machining skill.

The first failure mode is buying capacity before demand. A 5-axis center financed on a promise will not be paid for by a promise. If you cannot name the parts that will run on it, do not sign.

The second is quoting tight tolerance at loose-tolerance prices. A ±0.005 mm callout on a 200 mm part needs a controlled room, a finishing pass and inspection time. Charge for all three. Shops that win this work on price alone usually lose money on every repeat order.

The third is one-customer dependence. When 70 percent of revenue comes from one buyer, that buyer sets your prices and your payment terms. Aim for no single customer above 40 percent of monthly revenue by the end of year two.

The fourth is ignoring paperwork. Late delivery usually traces back to a missed process step, not a slow spindle. A simple traveler with sign-off at each operation catches most of it. Historical data from established shops shows that late delivery stays under 2 percent when every operation is signed off and inspected before the next one starts.