Most shops do not choose a cutting process part by part — they own one machine and quote everything through it. But when a job lands that your machine handles badly, or when you sub-contract a cut you cannot do in-house, the question of which process to run becomes a money question, not a technical one. Picking laser vs plasma vs waterjet correctly is the difference between a part that flows cleanly from the bed to shipping and one that needs three secondary operations to undo what the wrong process did to the edge. Get the choice right and the cut is cheap; get it wrong and you pay for it twice.

This guide compares the three thermal-and-cold cutting processes the way an estimator has to think about them: not "which is best" in the abstract, but which one produces the lowest finished cost for a given material, thickness, tolerance, and quantity. The headline machine rate is only half the story — edge quality and downstream rework decide the rest.

The three processes at a glance

Each process owns a band of work where it is clearly the cheapest sensible choice, and bleeds money outside it. The table below is the mental model; the sections after it explain how to apply it to a real drawing.

Factor	Laser (fibre)	Plasma	Abrasive waterjet
Sweet-spot thickness	0.5–20 mm	6–40 mm	10–150 mm+
Materials	Steel, stainless, aluminium, brass	Any conductive metal	Almost anything: metal, stone, glass, composite, titanium
Tolerance (typical)	±0.1 mm	±0.5–1.5 mm	±0.1–0.2 mm
Edge quality	Excellent, square	Bevel + dross	Clean, slightly tapered, no HAZ
Heat-affected zone	Small	Noticeable	None (cold cut)
Cut speed	Very fast (thin)	Fast (thick)	Slow
Running cost / hour	Low–medium	Low	High (garnet + pump wear)
Capital cost	High	Low–medium	High

Read the table as a set of trade-offs, not a ranking. A 2 mm stainless panel is a laser job every time; a 30 mm structural gusset is plasma; a 60 mm titanium plate or a heat-sensitive composite is waterjet — and quoting any of those on the wrong machine produces a number that is either uncompetitive or unprofitable.

Thickness and material: the first filter

Before edge quality or tolerance enters the conversation, thickness and material decide what is even on the table. This is the cheapest decision to get right because it is the most objective.

• Thin sheet, conductive metal: laser dominates 0.5–20 mm steel, stainless, and aluminium. It is fast, the kerf is narrow so material yield is good, and the edge needs little or no finishing. The detailed mechanics live in the laser cutting cost calculation guide.
• Thick plate, conductive metal: plasma takes over from roughly 6 mm upward, and through the 12–40 mm band it is far cheaper per hour than laser and faster than waterjet. It only cuts metals that conduct, and it leaves a bevelled, dross-bearing edge.
• Very thick, reflective, or non-conductive: waterjet is the only one of the three that does not care what the material is or how thick it gets within reason. Titanium, copper, stone, glass, carbon-fibre laminate, stacked sheets, 100 mm steel — all are routine waterjet work and impossible or ruinous on the other two.

Notice the overlaps. In the 6–20 mm steel band all three can physically cut the part, and that is exactly where the choice stops being obvious and starts being an estimating decision.

Edge quality and the hidden secondary-operation cost

The trap in cut-process quoting is comparing machine rates and ignoring what happens after the part leaves the bed. A cheap cut that needs grinding, a deburr pass, and bevel correction is not cheap — it just moved the cost downstream where the original quote never looked.

• Laser leaves a square, clean edge that is usually ready for the next operation. Minimal deburr, no bevel to correct, tight enough for most assemblies straight off the machine.
• Plasma leaves a few degrees of bevel and dross on the underside. If the drawing tolerance or a downstream weld prep needs a square edge, you are adding a grinding or machining operation that can cost more than the cut itself.
• Waterjet leaves a clean, cold edge with no heat-affected zone and no slag — critical when the part is heat-sensitive, will be welded to a tight spec, or has a tolerance the bevel of plasma would blow.

When the cut feeds a weldment, the edge condition matters even more, because weld prep and fit-up are where bad edges multiply labour — the welding and fabrication cost estimation guide covers how that downstream cost stacks up. The estimating rule is simple: cost the finished edge the drawing requires, not the raw cut.

A selection checklist you can run on any drawing

Run these five questions in order and the process usually picks itself. Stop at the first one that forces a decision.

1. Is the material non-conductive, reflective, or heat-sensitive? If yes — composite, stone, glass, copper, titanium where HAZ is forbidden — it is a waterjet job. Done.
2. Is the plate thicker than ~25 mm? If yes, it is plasma (conductive metal) or waterjet (everything else). Laser is out.
3. Is it thin conductive sheet under ~20 mm? Laser is almost certainly cheapest and cleanest — unless tolerance or HAZ rule it out.
4. Does the tolerance demand ±0.2 mm or tighter, or a perfectly square edge? That excludes plasma; choose laser (thin) or waterjet (thick).
5. What is the quantity? High volume rewards the fastest process (laser on thin, plasma on thick). One-offs and prototypes care less about cycle time and more about avoiding tooling and secondary ops.

The checklist is a filter, not a calculator. Once it narrows the field to one or two viable processes, you still build the actual cost the same way you build any cut quote — machine time at the burdened rate, material from nested yield, consumables, and amortised setup.

Worked example: the same bracket, three ways

Take a 12 mm mild-steel bracket, 300 × 200 mm, with eight holes and a 600 mm total cut length — squarely in the overlap band where all three can cut it. The numbers below are illustrative burdened-rate scenarios, not quotes; the point is the shape of the comparison, not the exact euros.

	Laser	Plasma	Waterjet
Burdened rate (€/h)	90	60	130
Cut + pierce time (min)	4.5	3.0	14
Machine cost / part	€6.75	€3.00	€30.30
Secondary edge work	None	Grind bevel (~€4)	None
Finished cut cost	~€6.75	~€7.00	~€30.30

The lesson is in the last two rows. Plasma wins on raw machine cost but the bevel-grind nearly erases the gap; laser ends up cheapest finished on this thickness because the edge is ready to use; waterjet is the most expensive here and would only be chosen if the part were heat-sensitive or needed a tolerance the others cannot hold. Move the same exercise to 30 mm plate and plasma pulls clearly ahead of laser, which slows dramatically. The right answer depends entirely on thickness, tolerance, and what the edge has to do next.

Putting it in the quote: rate, margin, and consistency

Whichever process wins, the price is built the same way: fully burdened machine rate × cycle time, plus nested material, plus consumables and assist media, plus amortised setup, then margin on the total. The burdened rate per machine — not a shop average — is the anchor, and the machine shop hourly rate calculation guide shows how to build it for each table. Apply margin with the proper formula (price = cost ÷ (1 − margin %)) as covered in margin vs markup, so the clean jobs are not quietly subsidising the messy ones.

The discipline that separates a profitable cutting shop from a busy one is consistency: the same process-selection logic and the same cost build on every drawing, so two estimators quoting the same part — or the same estimator on a Friday afternoon — land on the same number.

From drawing to the right process, automatically

The slow part of process selection is reading the drawing: pulling material, thickness, tolerance callouts, hole count, and cut length off the print before you can even decide which machine should run it. Under deadline pressure that reading gets shortcut, the tolerance gets missed, and a part that needed waterjet gets quoted — and cut — on plasma.

QuoteBuddy reads the technical drawing and surfaces exactly those inputs — material and thickness from the title block, geometry and cut length, hole and contour count, and the tolerance callouts that drive process choice — so the estimator decides from a complete picture instead of a hurried glance. It then builds the cost from your machine rates, cut speeds, consumable costs, and target margin, the same way every time, across laser, plasma, and waterjet alike.

Start a 30-day trial and run a few real drawings through it — from upload to a complete, itemised quote. See whether the process it points to, and the cost it builds, match what you would have done by feel, and where the wrong machine was quietly costing you margin.