Two numbers stamped on the underside of a pot, and nearly everyone reads them as a score. 18/10 sounds like it beats 18/8, which sounds like it beats 18/0, and a large amount of kettle and server pricing rests on exactly that reading. The numbers are real and they describe something precise. What they describe has almost nothing to do with how the object behaves once hot coffee is inside it.
18/10 means 18 percent chromium and 10 percent nickel by weight, with the balance mostly iron. It is a composition spec, not a quality rating. Chromium is the element that makes steel stainless. Nickel changes the crystal structure of the alloy, which buys a little extra corrosion resistance and a slightly warmer surface lustre. Both 18/8 and 18/10 belong to the grade 304 family, which is what most food-contact coffee equipment is made from, and we know of no evidence that a drinker can tell one from the other in the cup.
What the chromium is doing
Plain steel rusts because iron reacts with oxygen and the resulting oxide flakes away, exposing fresh iron underneath. The reaction never stops. It just keeps eating.
Add chromium above roughly 10.5 percent and the behaviour inverts. Chromium reacts with oxygen faster than iron does, and the chromium oxide it forms is dense and tightly bonded, and only nanometres thick. It seals the surface instead of flaking off it. Scratch it and it re-forms within seconds, provided there is oxygen available. That self-repairing film is the entire mechanism, and it is why the first number is the one that matters.
This is also why stainless is the correct answer for coffee at all. Rao and Fuller measured brewed coffee across several origins and published the numbers in Scientific Reports in 2018. It is a direct measurement study of brews rather than a trial of anything, and their hot brews ran pH 4.85 to 5.10, the Ethiopian sample most acidic and the Brazilian least. That is acidic enough to matter for some metals and not remotely acidic enough to breach a passive chromium layer. The FDA Food Code prohibits copper and copper alloys in contact with food below pH 6 and marks that provision a Priority item, the top of its three categories of importance (2022 edition, §4-101.14(A)). The provision names copper and copper alloys. It sets no equivalent use limitation on stainless.
What the nickel is doing, and why it costs more
Nickel does not make the steel more stainless in any direct sense. It changes what crystal structure the iron adopts as it cools.
With nickel at 8 to 10 percent, the alloy holds an austenitic structure at room temperature. That structure is non-magnetic, formable without cracking, and marginally better at resisting the kind of localised pitting that chlorides cause. It also takes a brighter polish and reads slightly warmer to the eye, which is a real part of why hollowware makers specify 18/10 for anything a guest will see.
Drop the nickel to zero and you get 18/0, which is grade 430: ferritic, magnetic, cheaper, and noticeably more willing to develop rust spots where water sits. Perfectly serviceable for a spoon. Less good for a vessel that holds a wet grounds bed for four minutes at a time.
There is a third case worth knowing. Some low-cost pots and flasks use 200-series steel, where manganese and nitrogen substitute for most of the nickel. It is a genuine stainless steel and it is a general property of the series that it pits more readily than 304 in chloride-heavy conditions, which in a kitchen means salt and dishwasher residue.
The magnet test, and the two ways it lies
Put a fridge magnet on the piece. Austenitic 304, which covers both 18/8 and 18/10, is essentially non-magnetic in its annealed state. Ferritic 430 grabs the magnet hard.
Two caveats stop this from being a clean verdict.
Cold working induces magnetism in 304. Deep-drawn and spun shapes, which is most kettle bodies and most cheap milk pitchers, can pick up a faint magnetic response at the rim or base without being anything other than 304. Faint is the operative word. A magnet that hangs on rather than clings is not evidence of a cheaper alloy.
And induction hobs need a magnetic base to work at all, so a great many good 304 pans have a ferritic disc bonded on deliberately. Test the wall, not the bottom.
The number nobody prints, which matters more
Gauge. Wall thickness. The grade tells you what the metal is made of and says nothing at all about how much of it there is.
Published thermal conductivity figures put copper near 400 watts per metre-kelvin, aluminium near 235, and austenitic stainless near 15. Stainless is a poor conductor by a wide margin, which cuts both ways depending on the job. A thin stainless server loses heat fast because it has almost no mass to hold it. A thin stainless pan develops hot spots because heat cannot spread sideways quickly enough, which is precisely why serious stainless cookware is clad around an aluminium or copper core.
For coffee, the practical version is short. Stainless is a good material for anything that needs to be inert, thin, cheap to form, and stable in a dishwasher: filter baskets, kettle bodies, cupping spoons, phin filters. It is a mediocre material for holding a drink hot unless it is doubled and evacuated, which is what a vacuum flask is. The decision about what pours and what holds is a different decision from the material, and pouring and holding are two jobs that want two different objects.
Kenneth Davids notes that the modern stove-top home roasters reviving a nineteenth-century design are fabricated in stainless where the originals were cast iron. That substitution is not cosmetic. Cast iron carries far more stored heat, and Davids traces a long arc in roasting technology away from beans contacting hot metal and toward beans suspended in moving air, precisely because metal contact scorches what touches it. The same physics governs your kitchen. Metal is an aggressive heat transfer surface, and how thick it is decides whether that is useful or a problem.
What stainless will not do
It will not stop coffee oils polymerising onto it. This is the single most common thing people blame on the metal, and the metal is innocent.
Metal filters let insoluble compounds and lipids straight through into the cup, which is why a French press has more body than a paper-filtered brew (Jessica Easto, How to Taste Coffee). Those same oils deposit on every metal surface they pass, dry into a brown varnish that a rinse does not touch, and eventually turn rancid. A press screen, a moka pot funnel or a portafilter basket left like this will make bitter, stale-tasting coffee out of perfectly good beans, and the owner will conclude that steel taints coffee. It does not. The film does. This is worth knowing before you swap gear: a metal filter changes the body of the cup rather than its strength, and the sludge in a French press is a technique problem before it is an equipment problem.
Stainless will also not fix a moka pot. The aluminium-versus-stainless argument in that category is mostly about durability and cleaning, not extraction, and neither metal turns a moka pot into an espresso machine. And 18/10 will not make a thin object feel substantial. Grade is not gauge.
The nickel question, stated honestly
Nickel comes up every time stainless is discussed, and the question people want answered is whether an austenitic vessel releases enough of it into hot coffee to matter to someone who reacts to nickel. We could not answer that from a peer-reviewed study or from an official body. We are not going to answer it from anywhere else.
So the honest position is no position. We will not tell you 18/10 is safe for nickel-sensitive people, and we will not tell you it is a risk. Both statements would be inventions. What we can tell you is that the 304 family is what hot beverage equipment is generally made from, and that a diagnosed nickel sensitivity is a question for a clinician rather than for a coffee shop. We took the same position on tin linings for copper, and for the same reason: we could find only interested parties saying it was fine.
Worth adding, from the other direction. Easto notes that metallic is a listed sensory subquality within bitterness, which means a coffee can taste metallic with no metal having entered it. If you suspect your steel is doing something, taste it against a second cup rather than judging one alone.
The 1922 footnote
William Ukers ranked the metals in All About Coffee by how little coffee attacked them: silver first, then aluminium, then Monel, then tin, with nickel and copper acceptable and iron to be avoided. Stainless is absent from the list entirely, because commercial stainless cookware barely existed when he compiled it. He also held that whatever you brewed in, the finished coffee belonged in earthenware or glass rather than in metal at all.
One line in that ranking is now not just dated but wrong to act on. Ukers put copper in the acceptable column. The FDA Food Code prohibits copper contact with food below pH 6, and brewed coffee sits below pH 6. Read the list as a record of what the trade believed in 1922, not as guidance about what to drink from.
A century of alloy development has otherwise overtaken the ranking. His conclusion has not. Stainless is now the default brewing metal, and a ceramic or glass vessel is still the better place to let the coffee sit.
What to actually check
Read the grade if it is printed, and treat 18/8 and 18/10 as the same answer. Weigh the object in your hand, because gauge is the variable the marketing copy omits. Test the wall with a magnet if the seller is vague, remembering that a bonded induction base is supposed to be magnetic. Then look at whether anything on it is not steel: a plastic gasket, a soldered seam, a bonded base disc, a lining.
That last one is where the grade genuinely decides something. On a copper vessel, the lining is the only thing standing between an acidic drink and a metal the Food Code prohibits from touching it. Stainless is one of the materials used for that job, alongside tin and nickel. We are not going to rank the three, because we looked for an official body that ranks them and did not find one. The question a lining has to answer is not which material is best. It is whether any copper is exposed. Our copper servers carry the lining material on the listing, and where a supplier will not confirm it in writing, we do not list the object at all.
Everything else in this category is a preference. The two numbers on the base tell you what the alloy is. How the object performs is decided by how thick it is, how it is joined, and how often it gets properly cleaned.