25 juil. 2025

Vegan Frozen Desserts: Nice Cream & Quick Sorbets

Two frozen bananas blended together give two servings at 108 kcal, 18.7 g of sugars and zero free sugars. The same weight of mango with two tablespoons of agave syrup gives a single bowl at 164 kcal, of which 13.6 g are free sugars, or 45 % of the daily limit the NHS recommends for an adult. Both come out of the same blender, in three minutes, made of fruit. The gap has nothing to do with the fruit. It comes from the syrup.

That is the point vegan frozen dessert recipes tend to skip, and it is also the easiest one to fix. The rest of this page weighs the classic recipes against the Swiss Food Composition Database published by the Federal Food Safety Office, explains why texture hinges on one measurable variable, and gives the safety marker that genuinely matters when you work with frozen berries.

What each classic recipe actually delivers

Every figure below comes from the Swiss Food Composition Database V7.1, applied to the exact quantities used in standard recipes. A peeled banana weighs roughly 120 g, so “two bananas” means 240 g of fruit.

Preparation	Serving	Energy	Sugars	of which free sugars	Fibre	Saturates
Banana-only nice cream	120 g	108 kcal	18.7 g	0 g	3.2 g	0.1 g
Chocolate nice cream (cocoa + 1 tbsp maple syrup)	135 g	154 kcal	24.6 g	5.9 g	4.7 g	0.7 g
Strawberry vanilla nice cream	195 g	136 kcal	22.9 g	0 g	6.1 g	0.1 g
Coconut pineapple nice cream (40 g coconut milk)	215 g	190 kcal	26.9 g	0 g	4.1 g	3.9 g
Peanut butter nice cream (32 g + 2 dates)	144 g	233 kcal	24.9 g	10.4 g (blended dates)	5.1 g	2.2 g
Mango sorbet (150 g + 2 tbsp agave syrup)	170 g	164 kcal	33.0 g	13.6 g	2.4 g	0.2 g
Mango sorbet without syrup	150 g	102 kcal	19.4 g	0 g	2.4 g	0.2 g
Red berry sorbet (+ 1 tbsp maple syrup)	170 g	116 kcal	20.0 g	11.7 g	6.1 g	0.1 g
Watermelon lime sorbet	160 g	90 kcal	18.6 g	6.8 g	0.8 g	0.2 g

Three things stand out. First, the energy gap between the lightest and the richest preparation runs from 90 to 233 kcal, a factor of 2.6, even though all of them go by the same name of vegan frozen dessert. Second, the syrup on its own decides the free sugars column. Third, coconut milk takes saturated fat from 0.1 to 3.9 g per serving, thirty-two times more, for an ingredient usually presented as a small creaminess boost.

Fruit sugars and free sugars: the distinction that changes everything

The NHS defines free sugars as all sugars added to food, plus those naturally present in honey, syrups, fruit juices and smoothies. Sugars in whole fruit and in milk are excluded. The recommendation for an adult is to stay under 30 g of free sugars a day, roughly seven sugar cubes.

The World Health Organization frames the same limit differently: under 10 % of total energy intake, about 50 g for someone eating 2000 kcal, with possible additional benefit below 5 %, or 25 g. Both institutions file maple syrup and agave syrup in exactly the same place as white sugar.

This is where the logic of vegan recipes turns against itself. The 240 g of banana in a nice cream supply 37.4 g of sugars and zero free sugars, because the fruit is whole, the cell matrix is intact and the fibre stays put. The tablespoon of maple syrup added “to sweeten it a bit” supplies 11.7 g, all of them free. The syrup accounts for about an eighth of the mixture by weight and delivers the entire problematic fraction.

A 2024 review in European Journal of Nutrition on how physiological responses to sugars vary by food source concludes that physical structure, energy density, fibre, potassium and polyphenol content explain part of the observed differences between whole fruit and juice, with measurable effects on glycaemic control and short-term appetite. That does not make blended fruit the equivalent of a bitten apple, but it clearly separates it from a syrup.

In practice one instruction covers it: taste before you sweeten. A properly ripe banana, a mango at its peak or strawberries in season need nothing added. If the fruit tastes flat, the problem is the fruit, not the missing syrup.

What Swiss law calls a sorbet

The federal ordinance on foodstuffs of plant origin (SR 817.022.17) devotes its chapter 4 to edible ice. Article 17 defines it as a frozen or semi-frozen preparation intended to be eaten frozen. Annex 3 then sets out what each name requires.

Denomination	Main requirements
Sorbet	Fruit at least 20 % by mass (10 % citrus, 6 % lemons), total dry matter at least 25 %
Water ice	Fat at most 3 %, total dry matter at least 15 %
Ice cream	Milk fat at least 8 %, total dry matter at least 30 %
Glace	Fat at least 3 %, non-fat milk solids at least 8 %, total dry matter at least 30 %

Article 18 paragraph 3 adds an interesting physical constraint: edible ice may not weigh less than 450 g per litre of finished product. A litre of commercial ice cream can therefore hold up to 55 % of air whipped in during freezing, which explains part of its softness and why a home preparation, blended without air, comes out of the freezer considerably harder.

The point that matters directly for the recipes on this page: the dry matter of fruit, calculated as 100 minus its water content in the Swiss database, sits between 9 and 25 %.

Fruit	Dry matter	Reaches the 25 % sorbet threshold
Banana	24.2 %	No, narrowly
Mango	18.9 %	No
Blueberry	14.3 %	No
Pineapple	13.6 %	No
Raspberry	13.2 %	No
Strawberry	9.7 %	No
Watermelon	9.0 %	No

No pure fruit purée clears the legal sorbet threshold. These rules apply to products placed on the market, not to what you make at home, and nobody is coming to check what you call your dessert. But they say something useful about the physics of the problem: commercial frozen desserts reach 25 to 34 % dry matter because sugars, fibre and fat lower the freezing point and stop water from forming large crystals. A watermelon purée at 9 % dry matter is essentially water, and it freezes like water.

Why texture depends on dry matter

A soft frozen dessert is one where the water sits in crystals small enough that the tongue cannot pick them out. Everything that is not water, meaning sugars, fibre, protein and fat, gets in the way of large crystal growth. The higher the dry matter, the more workable the mixture stays at low temperature.

That is why a banana nice cream works and a watermelon purée does not. At 24.2 % dry matter, banana is the densest fruit in solids on the fresh shelf, and its pectin adds a viscosity strawberry cannot match. At 9 % dry matter, watermelon gives you a granita, and no blender setting will change that.

Industry compensates with emulsifiers and stabilisers. A 2025 trial in Journal of Texture Studies on plant-based frozen desserts made from soy and pea protein isolates measured that mean ice crystal size was affected neither by emulsifier type, nor by its level, nor by stabiliser level, nor by protein source: these additives act on melting resistance and on partial coalescence of fat globules, not on crystal fineness. In other words, a home recipe cannot buy its way out of insufficient dry matter with one clever ingredient.

What you can do comes down to a few concrete levers. Choose dense fruit as the base: banana, mango, avocado. Add nut butter or coconut milk, which bring fat and dry matter at the same time. Serve straight after blending, when texture is at its best. And accept that a preparation stored three days in the freezer will come out hard, which is normal behaviour for a low dry matter mixture rather than a failure.

A 2026 trial in Foods shows the scope on the fat side: adding sunflower oleosomes at 24 % to a sugar-free plant-based ice cream improved emulsion stability and melting resistance, producing a measurably softer (30.74 ± 0.28 N) and less adhesive (1.87 ± 0.20 mJ) texture. The principle transfers to a home kitchen in a simple form: a little fat helps, provided you accept the calories that come with it.

Frozen berries: the safety point recipes skip

This is the only documented health risk in this type of preparation, and it is missing from almost every quick sorbet recipe.

In 2012, nearly 11 000 people, mostly children in German nurseries and schools, were caught up in the largest foodborne gastroenteritis outbreak ever recorded in that country. Investigations run by the federal government and the Länder identified a batch of frozen strawberries contaminated with norovirus, prepared in several catering kitchens. The Bundesinstitut für Risikobewertung updated its guidance sheet for institutional catering as a result.

The mechanism is straightforward: berries can be contaminated during irrigation, fertilisation, harvest or packing, including through water added during freezing. Cold does not destroy viruses. A 2026 review in Comprehensive Reviews in Food Science and Food Safety puts it plainly: freezing suppresses microbial growth but rarely eliminates micro-organisms, and the real risk plays out during thawing and the handling that follows.

The matter is not closed. In 2026, Emerging Infectious Diseases documented 24 hepatitis A cases in the Netherlands between November 2024 and February 2025, all part of one genotype IA cluster; consumption data and detection of the outbreak strain’s RNA in a food product pointed to frozen blueberries sold by a supermarket chain.

The BfR marker is core heating above 90 °C, judged capable of fully inactivating any viruses present. Quick microwave thawing is not enough, and neither is brief or uneven heating.

That marker is obviously incompatible with a quick sorbet, so the practical consequence is framed in terms of audience rather than temperature:

For a healthy adult, the risk attached to frozen berries of known origin remains low, and that is the trade-off most people make.
For young children, older people, pregnant women and immunocompromised people, the official recommendation is not to serve frozen berries without sufficient heat treatment. A sorbet for these groups is made with washed fresh fruit, with fruit you froze yourself after washing, or with berries cooked first, cooled and refrozen.
Mango, pineapple and banana, all peeled before freezing, do not fall into this risk category: the part you eat was never exposed.

One methodological detail is worth knowing if you read press coverage on this. A 2025 study in Food and Environmental Virology compared cell culture and qRT-PCR detection on frozen raspberries inoculated with murine norovirus: the limit of detection was 3.1 log PFU per 25 g in culture against 1.0 log PFU per 25 g by PCR, a two-log gap. A substantial share of genomes detected by PCR therefore does not correspond to infectious particles, and a positive test at a high Ct value does not necessarily mean danger.

Storage: the temperatures that matter

Recipes talk in days. Food safety authorities talk in degrees, and the distinction between safety and quality is useful here.

The UK Food Standards Agency puts the freezer at around -18 °C and the fridge between 0 and 5 °C. At -18 °C, a home frozen dessert does not become dangerous: bacterial growth is halted. What degrades is texture. Crystals grow through recrystallisation, the mixture turns grainy, and freezer burn, that surface drying caused by cold dry air, affects quality without affecting safety. The FSA puts optimal quality for frozen fruit and vegetables at 3 to 4 months.

Three practical rules follow. Pack airtight, in a shallow container rather than a deep one, to limit the surface exposed to air. Take the preparation out 10 to 15 minutes before serving rather than microwaving it, which would create melted patches that recrystallise into coarse grains. And do not refreeze a fully thawed dessert: the FSA only allows refreezing after cooking, with a single reheat afterwards.

If you freeze your own fruit, cut it into pieces and spread them on a tray before transferring them to a bag once hard. A compact 500 g block takes hours to freeze through, an isolated piece a few dozen minutes, and freezing speed directly determines crystal size within the fruit itself.

The method, no special equipment

For a nice cream, peel the bananas before freezing, which is impossible afterwards, and cut them into 1 to 2 cm slices. Freeze at least four hours, on a tray, with the slices not touching. Blend in pulses of a few seconds, scraping the sides between each: the mixture goes through a grainy stage that looks like a failure before turning creamy all at once. If the blender struggles, leave the slices out of the freezer for five minutes rather than adding liquid, which dilutes the dry matter you are trying to preserve.

For a quick sorbet, blend the frozen fruit with the minimum of liquid, one to two tablespoons per 150 g of fruit, and serve straight away. Optimal texture lasts a few minutes; past that, put it back in the freezer for twenty minutes rather than serving a lukewarm purée.

A stick blender works with small pieces and a patient method, but the result stays more uneven than with a jug blender. The limiting factor is motor torque, not speed.

Adapting to different profiles

For a child, the NHS free sugar thresholds fall with age: 24 g a day from 7 to 10, 19 g from 4 to 6, 14 g from 2 to 3, 10 g at one year. Below one year, no sugar should be added to food or drink at all. A pure banana nice cream therefore suits a three-year-old perfectly, while the mango sorbet with agave syrup, at 13.6 g of free sugars, uses up almost their entire daily allowance.

For anyone watching energy intake, the banana-only version at 108 kcal per serving compares well with a scoop of commercial dairy ice cream, which the Swiss database puts at 185 kcal per 100 g for the fruit variety and 278 kcal for the flavoured one. The chocolate-coated stick reaches 359 kcal and 15.2 g of saturated fat per 100 g, against 0.1 g for banana nice cream.

For an athlete or anyone after a filling snack, the peanut butter version brings 233 kcal and a little more protein, without turning the dessert into a protein source: you land around 5 to 6 g per serving, well short of a soy yoghurt.

For someone with diabetes or paying attention to their glycaemic response, the presence of fibre and the absence of syrup clearly separate a nice cream from a commercial sorbet, which the Swiss database puts at 32.1 g of sugars per 100 g. That replaces no individual monitoring, and a fruit-based preparation remains a sweet preparation.

For people with allergies, two concrete points: peanut butter recipes are out with a peanut allergy, including through blender cross-contamination, and coconut milk does not pose a problem for most people with tree nut allergies, though that question belongs to individual medical advice.

The mistakes that keep coming back

Adding syrup by reflex. It is the single decision in the whole recipe that moves free sugars from zero to a third or half of the daily limit. Taste first.

Using bananas that are not ripe enough. A green banana still holds starch that has not converted to sugars; it gives a bland, floury result that then gets corrected with syrup, closing the circle. Wait until the skin is speckled brown.

Adding liquid as soon as the blender struggles. Every spoonful of plant drink lowers dry matter and pushes the result towards granita. Let the fruit soften for five minutes instead.

Serving raw frozen berries to vulnerable groups. This is the only point on this page that concerns safety rather than comfort.

Storing in a deep container. Exposed surface drives freezer burn, and a thick block will not take a spoon after three days. Spread it to 3 to 4 cm thick at most.

What this page cannot claim

The nutritional figures are calculated, not measured in a laboratory on the preparations described. They rest on Swiss database averages, which do not capture variations in ripeness, variety and origin: two mangoes can differ by several grams of sugars per 100 g.

The Annex 3 thresholds apply to products placed on the Swiss market, not to home preparations. They are quoted here as a physical reference point, not as a constraint that binds you.

The data on viruses in frozen berries come from German and Dutch episodes. No Swiss data specific to contamination of retail frozen berries is published, and the real frequency of such contamination in batches sold in Switzerland is not quantified in accessible sources.

Finally, the texture trials cited concern industrial formulations with emulsifiers and stabilisers, tested under laboratory conditions. They illuminate the mechanism; they do not predict what a home blender will produce.

Frequently asked questions

Does a nice cream count as a portion of fruit? Yes in terms of fibre and micronutrients, since the whole fruit is retained. The WHO recommends at least 400 g of fruit and vegetables a day above age 10, and 120 g of banana contributes to that. The caveat is satiety: blended fruit goes down faster than fruit you chew.

Can you freeze a nice cream that has already been blended? Yes. It will set considerably harder than fresh out of the blender, because its dry matter is low. Take it out 10 to 15 minutes before serving. At -18 °C safety is not the issue; quality degrades over a few weeks.

My quick sorbet is too runny, what now? You added too much liquid while blending. Add more frozen fruit rather than putting the lot back in the freezer, which would give you a block to break up. The starting ratio is one to two tablespoons of liquid per 150 g of fruit.

Is agave syrup better than white sugar? The NHS and the WHO both classify them as free sugars. Agave syrup holds 68 g of sugars per 100 g in the Swiss database, maple syrup 58.6 g. The difference is about taste and fructose share, not nutritional status.

Do you need a powerful blender? No, but a high-torque motor helps. With a modest machine, cut the fruit smaller, work in pulses and let the fruit soften for a few minutes before you start.

Are these desserts suitable for children? A pure fruit nice cream is, allowing for the free sugar thresholds by age. For raw frozen berries, the official recommendation is to abstain with young children unless the fruit has been heat treated.

Can you make a nice cream without banana? Yes, with mango, which reaches 18.9 % dry matter, or with avocado, which brings fat. The result will be less creamy than banana and will need serving straight away.

How long does a home preparation keep? The FSA puts optimal quality for frozen fruit at 3 to 4 months. For a blended preparation, count on a few weeks before the texture turns distinctly grainy.

Going further

To dig into texture and crystals, our page on fruity ice creams without a machine sets out the mechanisms. On sweeteners, decoding sugar and its alternatives and the comparison of date, maple and lucuma sugars complete what is sketched here. If equipment is the obstacle, our criteria for choosing the right blender will spare you a few pointless purchases.

On desserts, sweet silken tofu recipes reach a creamy texture by another route, and festive desserts without refined sugar apply the same reasoning to other formats. For freezing in general, our anti-waste practices and our guide to storing fruit and vegetables properly give the durations and temperatures. Finally, matching recipes to what is actually in season remains the best way to avoid syrup altogether: see adapting meals to the seasons.