The person asking for their latte extra hot is not being difficult and is not ignorant. They have a real complaint, which is that the drink they paid for is lukewarm by the time they reach their desk, and they have proposed the obvious fix. The fix does not work. It solves a heat-loss problem by damaging the thing that was losing heat, and the damage happens in a narrow band of about five degrees, which is why the argument between the counter and the customer has run for twenty years without either side being able to point at a number.
Steam milk to roughly 60 to 65 degrees C and stop. That is where lactose and the milk's aromatics read at their sweetest and where the foam is still a single glossy fluid. Push past about 70 and whey proteins denature, sulphur compounds come off, and what reaches you is scalded milk rather than sweet milk, a change no technique reverses. Extra hot is not a stronger version of the drink. It is a different and worse one. If your coffee goes cold too fast, the variable to change is the cup, the drink size, or how long the cup sat empty on the shelf before it was filled.
What is actually happening in the pitcher
Steaming milk is two operations that share a wand, and they happen in a fixed order because the milk decides when the first one ends.
The first operation is stretching, which means introducing air. Air only folds into milk usefully while the milk is still cold. Once the base of the pitcher reaches roughly body temperature, around 37 degrees, any bubbles you add after that point stay bubbles. They sit on the surface as visible foam instead of dissolving into the body of the milk, and no amount of swirling afterwards repairs them. So the entire air budget of the drink is spent in the first few seconds, before the milk feels warm to the hand holding the jug.
The second operation is texturing, which is spinning the milk so the air already in it is broken into progressively smaller bubbles. How much air went in during the first few seconds is what decides whether you have made a cappuccino or a flat white. Texturing runs from about 37 degrees to the finish, and it is where the temperature question actually lives. The target is 60 to 65 degrees. Anette Moldvaer gives the tactile cue that beats any thermometer: hold your palm flat against the base of the pitcher and cut the steam at the moment your hand wants to come away. Racineux and Tran finish the same way, hand off the jug once it turns burning hot, and put the milk in the same 60 to 65 band.
Sound carries the rest of the information. Steam that shrieks is passing through the milk without catching, which is what low pressure sounds like, and no bubbles are being built. A steady tearing slurp is air going in. A low churning growl from the bottom of the jug means you are past boiling and the drink is finished. Volume tracks quality too: a jug of fine bubbles swallows more of its own noise than a jug of coarse ones, so a run that quietens as it goes is a run that is working.
Why 70 degrees is a wall and not a slope
Most cooking variables are gradients. You can be slightly over and get a slightly worse result. Milk temperature does not behave that way, which is the part that surprises people.
Whey proteins hold a folded shape. Heat past a threshold and they unfold, exposing sulphur-bearing groups that were previously tucked inside the molecule. Those released compounds are what your nose reads as boiled or faintly eggy. Because the unfolding is a threshold event rather than a steady drift, 75 degree milk does not taste like slightly warmer 60 degree milk. It tastes like a different ingredient. And the change is one-way. Letting it cool back down returns the temperature and not the flavour.
The foam goes at roughly the same point. Microfoam is a suspension held together by intact protein at the surface of every bubble. Denature enough of that protein and the structure will not form. Milk that has been boiled cannot be made into microfoam afterwards, however carefully it is treated, which is why a café that overheats a pitcher tips it and starts again rather than trying to rescue it.
The sweetness is manufactured, not added
The reason 60 to 65 is the target rather than simply a safe distance below the wall is that perceived sweetness peaks inside that window.
Nothing is added. The lactose in the milk was there when it came out of the fridge. What changes is how much of it your mouth registers, and how the aromatics released by gentle heating combine with it. Steaming is a flavour operation that happens to raise the temperature, not a heating operation that happens to affect flavour. That reframing is the whole disagreement in one sentence: the barista is treating the wand as a seasoning tool and the customer is treating it as a kettle.
Jessica Easto's account of the sensory research adds the other half. Heat at the top of the range costs you resolution, because a mouth busy with a near-scalding liquid registers less of what is dissolved in it. The Specialty Coffee Association's own cupping protocol will not let a professional taste a sample until it has fallen to 71 degrees, and holds back the assessment of sweetness until the cup approaches 38. So a drink served at 75 degrees is not only chemically worse, it arrives during the window in which your tongue is least able to detect what is left. You get the burn and not the sugar.
The customer's problem is real, and here is what actually fixes it
None of the above helps a person whose flat white is cold at the bottom of the cup. That complaint deserves a mechanical answer rather than a lecture.
Heat leaves a drink into the air across the top surface and into the vessel through the walls. The second one starts immediately and it is usually the larger loss for a small drink. A cold cup and the drink poured into it settle to a shared temperature within the first minute, and the heavier the cup the more heat it takes to get there. That happens before anybody has drunk anything, which is why rinsing the cup with hot water before you pour is the highest-value four seconds in the whole sequence.
Four things to ask for instead of extra hot:
- A preheated cup. Any café will do it. Most already do it for espresso and forget it for milk drinks.
- A smaller drink. A 470 millilitre latte spends fifteen minutes in your hand and loses heat the entire time. A 180 millilitre cappuccino is finished while it is still in the good band, and it is also a better-tasting drink for reasons that have nothing to do with temperature.
- A takeaway cup with a lid, even to drink in. Most of the surface loss is off the top. A lid is worth more degrees at the ten-minute mark than two extra degrees at the pitcher.
- Whole milk rather than skimmed, if what you want is a fuller drink. It will not keep the cup hot for longer and nobody should tell you it does. What it changes is body, because foam volume comes from protein and texture comes from fat.
There is one honest exception. If you are carrying the drink twenty minutes in a cold city and you know the last third will be the only part you drink sitting down, asking for it hotter is a rational trade of flavour for survivability. Say that out loud and most baristas will make it without complaint. The request that annoys people is not the heat, it is the implication that they have been doing it wrong.
The 65 degree number turns up somewhere else
This is a health-adjacent point and it is worth stating precisely rather than dramatically.
In 2016 the International Agency for Research on Cancer, the cancer agency of the World Health Organization, published Monographs Volume 116. Its working group classified the drinking of very hot beverages, defined as above 65 degrees C, as probably carcinogenic to humans, Group 2A. The cancer in question is squamous cell carcinoma of the oesophagus. The human evidence is observational, drawn mainly from case-control and cohort studies in populations that drink maté and tea at high temperatures, and IARC graded that human evidence as limited, which is its own term for an association that was observed but for which chance, bias and confounding could not be ruled out with reasonable confidence. That is an association. It is not a demonstration that hot drinks cause cancer.
The same working group graded the evidence on coffee itself inadequate and classified drinking coffee as not classifiable as to its carcinogenicity to humans, Group 3. That moved coffee out of the Group 2B possibly-carcinogenic listing it had been given in 1991, when a positive signal for bladder cancer in early studies was later judged to have possibly reflected other risk factors, tobacco smoking above all. Volume 116 is still IARC's standing evaluation. Checked August 2026.
We are not telling you that your latte is dangerous, and none of this is medical advice. Milk steamed to 60 to 65 degrees is already at or below the line IARC drew, and a cup only cools from there. The point is narrower: the temperature at which milk stops tasting good and the temperature at which a health body drew its line sit close enough together that "as hot as you can make it" has nothing at all recommending it.
What this will not do
Getting the milk temperature right will not fix the coffee underneath. Milk sits on top of espresso and modifies it. A shot that channelled, or beans four months past roast, will taste exactly as wrong at 62 degrees as at 72, with a slightly better texture. Diagnose the shot first.
It will not make a home machine steam like a café one either. The limiting factor on most domestic machines is steam pressure, not knowledge, and a wand that cannot spin the milk will produce large bubbles at any temperature you like. That is a purchase decision and no amount of technique substitutes for it.
And it will not settle the argument with the person who genuinely prefers a hotter drink. Preference is not a mistake. What the number does is tell them what they are paying for the extra heat, in sweetness and in texture, so the trade is made knowingly rather than by accident.
The single piece of equipment worth buying here is nothing. Put your hand flat on the side of the pitcher and take it off when you cannot keep it there. That is 60 to 65 degrees, it costs nothing, and it is what most professionals are using while the thermometer sits in a drawer.