Why Preheating an Air Fryer Heats the Basket

This page contains affiliate links. If you buy through one we may earn a commission, at no cost to you.

Both camps in that argument are right, because they are talking about two different things and neither has noticed. The air in an air fryer is not worth preheating and never was: all of it together holds less heat than a single chip. The metal is worth preheating, and the payoff is 174 kelvin at the surface the food is resting on.

So the honest answer depends entirely on whether the food touches the basket. For a tray of chips tumbling in moving air, skipping the preheat costs you almost nothing. For a steak or a salmon fillet lying flat on the crisper plate, it is the difference between landing on something that browns and landing on something that does not.

Weigh the air, because nobody ever does

Start with the quantity everybody assumes is the point. A Ninja 4 quart air fryer publishes the two figures the arithmetic needs, a 4 quart basket and a 1550 watt element, along with a temperature range of 105 to 400 Fahrenheit and a basket that holds 2 pounds of chips.

Four quarts is 3.7854 litres. At 400 Fahrenheit, which is 477.6 kelvin, air has a density of 0.7392 kilograms a cubic metre, so the basket contains 2.798 grams of air. Air holds 1,005 joules per kilogram per kelvin, so the entire air content of that machine has a heat capacity of 2.812 joules per kelvin.

Two point eight joules per kelvin. A single chip weighing five grams has a heat capacity of about 17 joules per kelvin. One chip is six times the thermal substance of all the air in the basket.

Scale it up to the biggest machine in our catalogue and nothing improves. An Instant Pot 6 quart Vortex Plus publishes 6 quarts and 1700 watts, which is 5.6781 litres holding 4.197 grams of air and 4.218 joules per kelvin. One chip still beats it four times over.

That single comparison settles the mechanism. The air is not a store of heat that you charge up and then spend. There is nothing in it to charge. It is a conveyor belt, and the only question about a conveyor belt is how fast it runs.

How fast it runs: eight to thirteen times a second

Work it from the published wattages. If the element puts 1550 watts into the air and the air gives up 50 kelvin of it as it passes the food, the mass flow has to be 1550 divided by 1005 times 50, which is 0.0308 kilograms a second. At that density it is 88.4 cubic feet a minute.

Now divide by the cavity. Eighty eight cubic feet a minute through 3.7854 litres is 11.0 complete air changes every second. The Instant Pot's 1700 watts through 6 quarts gives 8.1 a second. A CHEFMAN 2 quart mini air fryer, which publishes 900 watts and 2 quarts, gives 12.8.

Every air fryer in this catalogue replaces the air around your food between eight and thirteen times a second. That is the fan noise, that is why a crowded basket ruins the result, and that is why the crisper plate is perforated. Our page on using air fryer liners without ruining airflow is about protecting that figure, and the arithmetic says why a solid liner is so destructive: it does not slow the air a little, it removes the only delivery mechanism the machine has.

Put the whole job next to it. The Ninja's published 2 pounds of chips is 907 grams, which needs 293 kilojoules to get from fridge temperature to 100 degrees and perhaps another 164 to drive the surface water off, so 457 kilojoules in total, and at 1550 watts that is 295 seconds of pure energy delivery. Each complete pass of the air carries 141 joules. So the cavity has to be emptied and refilled roughly 3,250 times to cook one basket of chips. Preheating buys you one of those 3,250.

The preheat is the metal, by a factor of 178

If the air is not what you are heating, what is? The basket and the crisper plate.

No air fryer we carry publishes a basket mass, so take a representative 0.55 kilograms of aluminium for a 4 quart basket with its plate, which is 495 joules per kelvin. Bringing that from a 20 degree kitchen to 204 degrees is 184 kelvin, which is 91.1 kilojoules, and at 1550 watts it takes 58.8 seconds.

Bringing all the air through the same 184 kelvin takes 517 joules, which is 0.33 seconds.

A ratio of 178. A preheat cycle is almost exactly a metal-heating cycle with the air coming along for free, and that also disposes of the other half of the folk wisdom. When you pull the drawer out at the halfway point to shake it, the temperature you let out is 0.5 kilojoules. It is back in a third of a second. What you let out that genuinely matters is nothing, provided the basket stays in your hand rather than on the counter, because the basket is the part carrying 91 kilojoules.

Why 58.8 seconds of metal buys 174 kelvin

Now the part that makes preheating worth doing at all, and it is not about temperature in the cavity. It is about the instant of contact.

When two materials touch, neither of them wins. The interface settles at a temperature weighted by how readily each side can supply heat to the junction, and that property is thermal effusivity, the square root of conductivity times density times specific heat. Our page on why a Dutch oven bakes better bread works that method through properly and this page borrows it rather than re-deriving it.

Aluminium comes out at 23,998 in the usual units. A water-rich potato comes out at 1,401. The ratio is 17.1, so the metal almost entirely dictates the interface and the food barely gets a vote.

A crisper plate preheated to 204 degrees, meeting food at 5, gives an interface of 193.0 degrees. The same plate cold at 20 degrees gives 19.2. The difference is 173.8 kelvin, and the significance of that span is in our own page on why an air fryer browns the crust before the middle cooks, which establishes that browning needs 140 to 170 degrees. One hundred and ninety three is comfortably inside it. Nineteen is room temperature.

And the plate holds it. Its own 135 joules per kelvin falling from 204 to 150 gives up 7.29 kilojoules, which a 1550 watt element replaces in 4.7 seconds, so the element refills the plate faster than the food can drain it.

Which is precisely the distinction our own air fryer cooking times page already draws in words, and it draws it correctly: "Air fryers preheat quickly, often in just a few minutes, so long warm-up times are rarely needed. Many foods can even go into a cold basket, though preheating helps foods that benefit from an immediate sear, like a steak or a piece of salmon you want crisp on the outside." Every clause of that is right. One hundred and seventy four kelvin is the size of "helps".

One of our records makes the claim backwards

Four air fryer records in this catalogue take positions on preheating and they are not compatible with each other.

The Ninja 4 quart says "A 1550-watt supply keeps preheating quick and results consistent". A Ninja Air Fryer Pro AF141 at 5 quarts and 1750 watts says "A 1750-watt heating element drives rapid preheating and consistent cooking", and publishes a usefully concrete load of 4 pounds of chips or 5 pounds of wings. A Cosori Air Fryer Pro LE at 5 quarts ships a dedicated preheat program, with "Presets include preheat and keep warm". And the Instant Pot says the opposite: "Little to no preheating is a plus, as the specs note."

Three of those four are defensible. The Instant Pot's is not, and the arithmetic says so twice.

First, power density. Of the four machines that publish both a wattage and a capacity, the ladder runs 475.5 watts per litre for the 2 quart mini, 409.5 for the Ninja 4 quart, 369.8 for the Ninja 5 quart Pro and 299.4 for the Instant Pot 6 quart. It is monotonic, and the Instant Pot is at the bottom. It has the least power for every litre it has to heat, so it is the slowest of the four to bring anything up to temperature, not the fastest.

Second, the metal. Scale a basket's mass by volume to the two thirds power and the Instant Pot's 6 quart basket is about 0.72 kilograms, or 649 joules per kelvin, so its metal preheat is 119.4 kilojoules, which at 1700 watts takes 70.2 seconds. That is 19 percent longer than the 4 quart Ninja's 58.8 seconds. The machine marketing little to no preheating is the one with the longest preheat of the four.

What the claim is probably reaching for is the true and separate fact that an air fryer reaches temperature far faster than a full-size oven, which it does, and our page on air fryer versus convection oven covers the comparison. That is a statement about ovens and not a statement about this machine against its own shelf.

The lowest wattage in the group is the highest power density

The mini is the opposite case and its own record undersells it in the same way, from the other direction. Its drawbacks say "900W power may be lower than larger models", which is true in watts and misleading in the quantity that decides anything.

Nine hundred watts over 1.8927 litres is 475.5 watts per litre, which is the highest figure in the catalogue and 58.8 percent more than the Instant Pot's. That is why its headline claim holds up: a conventional oven running 2,500 watts over a 60 litre cavity is 41.7 watts per litre, so this machine carries 11.4 times the power density, and "cooks food up to three times faster than a conventional oven" is a conservative reading of an eleven-fold advantage in heating rate.

It is the same mistake our own records make about pressure cookers, and we have written it up before: comparing raw wattages across different capacities tells you about the plug, not about the cooking. Our page on why a slow cooker Low setting is a wattage runs into exactly this and finds a record drawing the comparison in the wrong direction for the same reason.

The product whose premise depends on a number it does not publish

A Ninja DZ201 Foodi 8 quart DualZone with two baskets is the most interesting record in the group because of what is missing from it. It publishes two independent 4 quart zones, 105 to 450 Fahrenheit, Smart Finish and Match Cook, and this specification of the hardware: each zone carries "its own basket, fan, and heater".

It publishes no wattage anywhere.

That matters more here than on any single-basket machine, because two baskets is twice the metal. Two sets of basket and plate is about 990 joules per kelvin, needing 182 kilojoules to preheat. If the pair shares one ordinary household circuit at around 1700 watts total, then each zone is drawing 850 watts and each basket takes 107 seconds to come up, against 58.8 for a single 1550 watt basket. Nearly twice as long, per zone.

Whether that is what happens is not knowable from the record, and that is the point. A product sold on the premise of two independent heated zones does not state the power those zones divide, which is the one figure that decides whether they are genuinely independent or merely separate. Its own drawbacks worry about the right things in the wrong units, flagging that it is "Large and bulky on the counter" and that "Two baskets mean more parts to wash". The physical question is whether two zones at once are each as fast as one zone alone, and nothing on the listing answers it. Our dual basket air fryer roundup has the rest of the category and runs into the same silence.

The honest limit, and the test worth doing once

Three things on this page are exact. The heat capacity of the air, which is the ideal gas density at a published temperature times a published volume. The contact temperature, which is the standard effusivity weighting, and which is insensitive to what the food is because the ratio between metal and food is 17.1 either way. And the air change rates, which are each machine's own published wattage divided by what its own air can carry.

What is representative rather than published is the metal, and it carries the headline figure. No air fryer we carry publishes a basket mass, a crisper plate mass, a fan airflow or an internal dimension, so the 0.55 kilograms of aluminium is an estimate from plausible wall thickness. Halve it and the preheat falls from 58.8 seconds to 29.4 and the ratio against the air from 178 to 89. Double it and the preheat is nearly two minutes. The conclusion is the same at every value, because the air is three orders of magnitude away and no plausible basket closes that.

The 50 kelvin drop across the food is also an assumption and it is the one input the air change figures are directly proportional to. At 25 kelvin every rate doubles, to 16 to 26 changes a second. At 100 it halves. None of those readings changes the answer to the question, because the air change rate is the thing that makes preheating the air pointless in the first place.

What this page deliberately does not compute is how long any given food needs, because that is a conduction problem through the food and our own browning page already owns it, including the floor it sets that no machine can beat. The arithmetic here stops at the surface.

Which leaves one test worth doing once on your own machine, and it takes two minutes. Set it running empty for a minute, then switch it off, pull the basket out and touch the crisper plate with the back of a spoon. That is the 91 kilojoules. Then do the same without the warm-up. Whatever you can feel between those two spoons is the entire thing preheating buys, and now you know which of your own recipes cares about it: anything lying flat on the plate, and nothing that tumbles.