For runs of between ten and a few hundred parts, additive and subtractive manufacturing genuinely compete, and the right answer depends on factors that are easy to assess once you know what to look at. The decision is often made on habit or on whichever supplier was asked first, which is how businesses end up paying more than they needed to or receiving parts that do not perform.
Start with the material requirement
If the part must be a specific engineering material with certified properties — aluminium 6061, stainless 316, acetal, PEEK to a specification — machining from stock gives you known, isotropic material behaviour. A machined part has the properties of the bar it came from.
Printed parts, particularly from FDM, are anisotropic: meaningfully weaker across the layer lines than along them. For many applications this is irrelevant. For a load-bearing bracket in a known direction, it is the whole question.
Where the requirement is a functional prototype or a low-stress component, the material argument largely disappears and other factors decide.
Geometry decides more than people expect
This is where additive wins outright. Internal channels, lattice structures, enclosed voids and organic forms are either impossible to machine or require multi-part construction and assembly.
Conversely, machining wins on simple prismatic parts, precise flat faces, accurate holes and threads, and anything requiring a fine surface finish as-produced.
A useful test: if you can describe the part as a block with material removed, machining is likely cheaper. If it looks like it grew rather than being cut, printing is likely cheaper.
Tolerance and finish
Machining routinely achieves tolerances an order of magnitude tighter than most printing processes, with surface finishes that need no additional work.
Printed parts typically need post-processing to reach a comparable finish, and that labour is frequently underestimated in the comparison. A printed part that needs sanding, priming and painting may cost more in total than a machined one that came off the mill ready.
Where tight tolerance is needed only at specific interfaces, a hybrid approach works well: print the part and machine the critical features afterwards.
How cost behaves with quantity
The cost curves differ in shape, and this is the crux of the decision.
Printing has essentially no setup cost and a per-part cost driven by material and machine time. The tenth part costs nearly the same as the first. Machining carries real setup — programming, fixturing, tooling — amortised across the run, so per-part cost falls steeply with quantity.
The crossover for typical small parts often sits somewhere between twenty and a hundred units, though it varies enormously with geometry. Below it, printing wins on cost; above it, machining usually does. The only way to know for your part is to get both quoted, which is worth doing whenever a run sits in that band.
Lead time
Printing is generally faster to first part, because there is no programming or fixturing. For iterative development, where you want three versions in a fortnight, this advantage compounds and often outweighs a per-part cost difference.
For a settled design in quantity, machining lead times are competitive and sometimes better, since a machine can produce parts considerably faster than a printer once running.
The realistic answer for product development
Most hardware projects use both, in sequence. Print during design iteration, where speed and geometric freedom matter and quantities are tiny. Move to machining or moulding as the design settles and volumes rise.
Companies moving through this transition tend to find the constraints are not only manufacturing ones — file management, version control, supplier documentation and quality records all become real problems at the same time. The infrastructure side of moving from prototype to product is worth planning alongside the fabrication decision, because discovering it late is expensive.
A practical approach
Quote both for anything in the ambiguous range. Supply the same file, state the same requirements, and compare total cost including finishing. Suppliers will often tell you honestly when their process is the wrong one, and that advice is worth having.