Is CNC Machining Additive or Subtractive Manufacturing?

• 1
  CNC Machining Additive or Subtractive Manufacturing: The Short Answer
• 2
  How Subtractive Cutting Removes Material
• 3
  How Additive Processes Build Layer by Layer
• 4
  Where Each Process Stops Working Well
• 5
  Why the Real Answer Is Often Both Processes
• 6
  CNC Machining vs Additive Manufacturing: Practical Comparison

CNC machining is subtractive manufacturing. The cutting tool follows a programmed path and removes material from a workpiece until the remaining geometry matches the CAD model. Milling, turning, drilling, boring, tapping and EDM all work this way. The starting point is a solid billet, bar, plate or casting, never a loose powder or a resin vat.

The control system is what makes it CNC. A post-processor converts CAM toolpaths into G-code, and the machine executes those coordinates without an operator turning handwheels. That is why two parts cut from the same program on different days land within a few micrometres of each other. The process is repeatable because the motion is digital, not because the machinist is careful.

Additive manufacturing builds the opposite way. A 3D model is sliced into layers, and material is deposited, fused or cured one layer at a time. Powder bed fusion, directed energy deposition, material extrusion and vat photopolymerisation all belong here. The part grows upward. Nothing is removed except support structures.

So the answer to the title question takes one sentence. CNC machining is subtractive. Additive is a separate family of processes that happens to use the same digital model as its input. The two share CAD files, not physics.

In subtractive machining, a sharp edge shears metal. The tool rotates or the workpiece turns, and the feed rate pushes the edge into the material at a controlled depth of cut. Chips form, heat leaves with them, and the surface left behind is the new part surface. Tool geometry decides what shapes are reachable.

Rigidity sets the limit on accuracy. Every pass generates cutting force, and that force deflects the tool, the holder and the workpiece. A long, thin end mill in aluminium will chatter before it reaches ±0.005 mm. A short, stubby tool in the same material holds tolerance easily. This is why part design and workholding matter as much as the machine.

Thermal behaviour also matters. Roughing removes most of the stock and puts heat into the part. If you finish immediately, the part is still warm and will shrink as it cools. For tight work we rough, let the part stabilise, then take light finishing passes. That sequence is standard for aluminium and stainless alike.

Some geometries are impractical to cut. Deep narrow slots, internal channels with bends, and hollow sections with no tool access are the classic examples. A 3 mm cutter can only reach about 12–15 mm deep before deflection ruins the finish. Beyond that ratio, additive or a different process is usually the better call.

Additive manufacturing starts with a sliced model. Each layer is a 2D contour, and the machine traces it before indexing upward by the layer thickness. Metal powder bed systems use a laser or electron beam to melt powder, typically in 20–60 μm layers. Resin systems cure a photopolymer with UV light.

The physics favours internal complexity. Because material is added rather than reached by a tool, conformal cooling channels and lattice structures are straightforward. A heat exchanger with curved internal passages is hard to mill and easy to print. That is the real advantage of additive, not raw dimensional accuracy.

The trade-off is surface and structure. As-built surfaces are rough, often Ra 8–15 μm, and stepped on sloped faces. Porosity, residual stress and anisotropic grain structure are common in metal parts. Most functional additive parts need stress relief, support removal and finish machining on the critical faces.

Shrinkage and distortion accumulate as layers stack. Long thin walls warp. Tall builds can drift. For that reason, additive parts that carry load or seal against fluid usually get their mating surfaces turned or milled afterward. The printed blank is a near-net shape, not a finished part.

Subtractive machining gets expensive when material removal is mostly waste. A bracket machined from a 200 × 200 × 100 mm aluminium block may turn 85% of the billet into chips. Additive would use far less material, but the printed part still needs finishing. The break-even sits around complex internal geometry, not part size.

Additive gets expensive when you need many identical parts. Machine time per part is long, powder handling adds cost, and post-processing is manual. Subtractive wins decisively at volumes above a few hundred units, and even earlier if the geometry is simple and the tolerance is tight.

Surface finish is another clean boundary. Fine machining reaches Ra 0.2–0.8 μm on aluminium and stainless. As-built additive does not come close. If a face is a bearing seat, a seal groove or an optical mount, it will be machined regardless of how the rest of the part is made.

Hardness and material choice cut both ways. Subtractive handles hardened tool steel, Inconel and titanium with the right tooling and speeds. Additive can make some alloys that are difficult to cast, but the available powder library is narrower than the bar and plate library. Check material availability before committing.

Most production parts that use additive are hybrid. The blank is printed close to shape, then the critical features are machined. This is common in aerospace brackets and medical implants, where internal lattice saves weight and machined faces carry the load. Each process does what it is good at.

The sequence matters. Print, stress relieve, then machine. Skipping stress relief means the part moves after machining and the tolerances open up. For titanium and Inconel, a full anneal between print and cut is normal practice, not an optional step.

At GreatLight we run the subtractive side of that chain. We machine printed blanks, castings and forgings to final tolerance on 3-axis, 4-axis and simultaneous 5-axis centers. We also machine from solid when that is the cheaper route. The decision is made per part, not per policy.

If you are unsure which route fits your geometry, send the STEP file. We will tell you whether the part is a good subtractive candidate or whether printing the blank first saves money. That answer comes back with the quote.