Every conversation about cutting discs starts the same way. Thin saves material. It does. The question is how much, and whether that is the number that should decide the order.
Take a fifty millimetre square bar and cut it through. What leaves as dust is the kerf: thickness, times depth, times the length of the cut. At 1.2 mm that is three thousand cubic millimetres. At 2.0 mm it is five thousand. At 3.0 mm it is seven and a half thousand. Steel weighs 7.85 grams per cubic centimetre, so those three numbers are 23.6 grams, 39.3 grams and 58.9 grams.
Going from a 3.0 to a 1.2 saves thirty-five grams a cut. Quoted on its own, that sounds like something. Put it beside the part it came off. A fifty by fifty by one hundred block is 250 cubic centimetres, and the 4.5 cubic centimetres you save is under two percent of it. Cut a ten metre bar into a hundred pieces and you have saved 3.5 kilograms out of 196. One and a half percent of the steel, bought with a disc that has six percent of the bending stiffness.
So when does kerf actually pay? Three cases, and they are narrow. The material is expensive enough that grams matter: titanium, tool steel, some stainless. The section is large enough that a single cut removes a real share of a billet. Or the cut sets a finished length, and kerf width is a dimension rather than waste. Outside those three, kerf is a rounding error wearing the clothes of a cost saving.
Cut to length is the case where kerf stops being waste and becomes arithmetic you cannot argue with. Forty pieces at a hundred millimetres, cut from one bar: the stock you need is forty lengths plus thirty-nine kerfs. At 3.0 mm that is 117 millimetres of kerf. At 1.2 mm it is 47. Seventy millimetres of bar, on a batch of forty, decides whether the last piece comes out of this bar or out of the next one. That is the case the thin disc was made for, and it has nothing to do with the weight of the dust.
What a thin disc really buys is not the material. It is the material per second. At the same feed, a 1.2 kerf means the disc removes forty percent of the volume a 3.0 disc has to, and the tangential force scales with that. So does the power the motor has to deliver, and so does everything downstream of the motor.
A thick disc on a 230 mm hand-held grinder does not cut slowly because the abrasive is worse. It cuts slowly because it asks for more torque than the motor has, the spindle sags, the rim speed drops, and the disc starts rubbing where it was cutting. The operator feels a disc that has gone soft and leans in. The cut wanders. The disc pinches in its own kerf. The report afterwards says the disc was no good, and the disc was doing exactly what the arithmetic said it would do.
There is a second effect running the other way, and it is the part most people miss. Contact area is kerf width times contact length, so a thin disc puts the same force onto less area. More pressure per square millimetre tends to keep a bonded disc open and self-sharpening, which is why narrow cuts often look cleaner than wide ones. But it also concentrates heat on a narrower line. Thin discs tend to run cooler in the body and hotter at the edge, and the wear you see is a narrow band at the rim rather than a wide one across the face.
Then the other side of the ledger, because thin is not free either. Bending stiffness goes with the cube of the thickness. A 3.0 mm disc is nearly sixteen times stiffer than a 1.2. In a deep cut, in a stack, on a surface that is not flat, the thin disc is the one that wanders and the one that pinches. If the cut is deep and the clamping is poor, the material you saved gets paid back in broken discs, and a broken disc is the expensive kind of failure because it arrives as a stoppage, not as a cost line.
This is why the catalogue carries a ladder instead of one thickness. 107, 115 and 125 at 1.2. 125 at 2.0. 180 at 1.6. 230 at 2.0 and 3.0. The ladder is not indecision. It is the acknowledgement that the right thickness is set by the cut, and that "thin is better" is only true until one of the four terms changes.
Ask the four questions in order. How expensive is the material. How deep is the cut. How well is the part held. How much power does the machine actually have. Thickness is the answer to those, not a preference and not a habit.
One caution on all of this. The numbers above use nominal thickness, and a real kerf is wider than nominal. Flange runout, a little side pressure, a worn flange: all three widen the slot. If the decision is close, cut once and measure the slot with a caliper. The arithmetic tells you which way the decision leans. The caliper tells you by how much.
One more thing worth saying plainly, because it comes up in every quotation. A cutting disc is not a grinding wheel that was made thinner. The 6 mm version at the same diameter is a different tool with a different job: it is built to be pushed sideways and to have wear reserve. Asking a 1.2 to do that work is not a thickness choice, it is a category error, and it fails in a different way than a thin disc normally fails.
So buy thin for the cut, not for the swarf. If the saving is under two percent of the part, you are not buying material, you are buying speed, and you should say so in the meeting. It is a better argument and it happens to be the true one.