Starch vs Fat Nut Pastes: Two Ways to Build Gelato Body
Table of contents
Every nut gelato needs body, and there are exactly two ways to get it. You can let the nut's own fat do the work, or you can build structure out of carbohydrate. The two routes cost different money, behave differently in the freezer, and taste nothing alike.
Two Different Jobs
Body in gelato is a perception, not a measurement, and two entirely different physical mechanisms produce it. Fat coats the palate and slows the release of everything else, which reads as richness and length. Hydrated starch thickens the unfrozen water phase, which reads as thickness and resistance to melting. A tub can be dense and lean, or rich and thin, and the difference is which lever you pulled.
The confusion starts because the market sells both under the same word. A jar labelled "hazelnut paste" might be 100% ground hazelnut, or it might be a formulation stretched with sugar, vegetable oil and starch or flour. Those are not two grades of the same product. They are two different tools.
What a 100% Nut Paste Brings
A pure paste is just the nut, roasted and milled until the cell walls give up their oil. Its composition is therefore the nut's composition, and USDA FoodData Central gives you the whole picture.
| Per 100 g (raw nut) | Fat | Protein | Carbohydrate | Water |
|---|---|---|---|---|
| Hazelnut | 60.8 g | 15.0 g | 16.7 g | 5.3 g |
| Almond | 49.9 g | 21.2 g | 21.6 g | 4.4 g |
| Pistachio | 45.3 g | 20.2 g | 27.2 g | 4.4 g |
Two things fall out of that table. First, a nut paste is a solids bomb: around 95 g of every 100 g is dry matter, so it moves total solids hard and fast. Second, the dominant solid is fat, which is why paste dosage is capped by the fat budget rather than by flavour.
At 10% of a milk base, hazelnut paste adds roughly 6 g of fat per 100 g of mix on its own, before any cream. That is enough to push a base straight past the upper end of the ideal fat percentage if the dairy was not adjusted down first. Pistachio paste is gentler on fat and heavier on protein, which is part of why pistachio bases so often feel drier than hazelnut at the same dose.
Quick reference. Route A adds body as fat: 100 g of hazelnut paste in a 1,000 g mix contributes about 95 g of solids of which 61 g is fat, and zero PAC. Route B adds body as hydrated carbohydrate: 15 g of cornstarch contributes 15 g of solids, no fat, and also zero PAC.
What Starch Brings
Starch does its work by gelatinising: granules absorb water, swell, and burst, releasing polymer chains that thicken whatever is left liquid. Maize starch does this at roughly 62 to 72 °C, so any base pasteurised at 85 °C has already taken it well past the transition. Below that window the granules stay intact and you get chalk instead of structure.
The payoff is water management. Every gram of water immobilised in a starch network is a gram that cannot migrate and refreeze into a large crystal during storage, which is why starch-thickened bases hold up so well against temperature abuse. Cornstarch in gelato covers the working dosages, and the mechanism sits alongside gums rather than competing with them, as starch vs gum sets out.
The cost is the amylose. Ordinary maize starch is roughly a quarter amylose, and amylose recrystallises as the product sits cold. That is retrogradation, and it shows up as a short, slightly grainy texture around day four. Waxy maize starch is effectively all amylopectin and largely dodges the problem, which is why it is the sensible default when the tub has to survive a week in the case.
Fat vs Starch, Side by Side
Neither route touches the freezing curve. Fat is not dissolved in the water phase and gelatinised starch is far too large a molecule to depress a freezing point, so both contribute essentially zero PAC. This is genuinely useful: you can add body without making the tub harder or softer, and tune texture and scoopability as independent problems.
Everything downstream of that, though, diverges.
Fat carries aroma. Most of the volatile compounds that make a roasted nut smell like a roasted nut are lipophilic, so they live in the fat phase and release slowly as it warms in the mouth. That slow release is the long finish people pay for. Push it too far and the same mechanism works against you: an over-fatted base mutes flavour rather than amplifying it, and the texture turns waxy at serving temperature.
Starch carries nothing. It is flavour-neutral by design, and a base leaning on it needs its aroma from somewhere else. What starch gives instead is cheap solids and melt resistance, plus a very forgiving response to imperfect freezer discipline. What it takes is definition; past about 2% of the mix, thickness starts reading as gumminess and the product feels pasty rather than dense.
Dosage, the Legal Floor, and Which Route to Pick
Working ranges are narrow on both sides. Pure nut paste sits at 8% to 12% of the mix for hazelnut and pistachio, with the dairy fat pulled back to compensate. Starch sits at 0.5% to 2%, and the top of that range is already firmly in gummy territory for a milk base. Neither range has much room in it, which is the practical reason the two routes get combined rather than chosen between: 9% paste with 0.8% waxy starch is a far more common professional answer than 12% paste alone.
There is also a labelling angle in the United States. 21 CFR 135.110 sets the floor for a product sold as ice cream at 10% milkfat and 20% total milk solids. Nut fat is not milkfat, so a hazelnut paste does nothing for that number. The standard does allow relief when bulky flavouring ingredients are used, permitting milkfat down to 8% and total milk solids down to 16% — but the relief is proportional to the bulky ingredient, not a blanket exemption, and it does not convert nut fat into milk fat. Plan the dairy block first and let the paste sit on top of a base that already complies.
Use the fat route when the nut is the product. Hazelnut, pistachio and almond paste flavours are bought for aroma intensity and length, and only the fat phase delivers those. Accept the cost per litre, adjust the dairy down, and resist the temptation to claw the richness back with cream you have already spent.
Use the starch route when the nut is a supporting character, when the base has to hold texture through a long display cycle, or when the fat budget is already spent by chocolate or by cream. A modest amount of starch under a modest amount of paste is often better than a large amount of either, and it is almost always the cheaper of the two per litre of finished product.
Where It Goes Wrong
The base is rich but the flavour is flat. Too much fat, not too little paste. Adding more paste makes it worse. Pull the cream down and re-taste before touching the paste.
It is thick in the machine and short on the spoon. Classic ordinary-maize retrogradation. Switch to a waxy starch and the same dose behaves.
The paste seizes when it hits the mix. A fat-continuous paste dropped into cold liquid will ball up. Temper it with a little of the warm base first, then return it.
The pistachio version feels dry next to the hazelnut version. It should. Pistachio carries about 15 g less fat and 5 g more protein per 100 g, so at equal dosage it is a leaner base. Compensate with dairy fat, not with more paste.
The starch dose fixed the texture and killed the aroma. Starch thickens the same unfrozen phase the volatiles have to travel through, so heavy dosing slows aroma release as surely as excess fat does. If the flavour went quiet the week you added starch, halve the starch before you question the paste.
Related Concepts
- Almond flour vs almond paste — the same fork, one ingredient down
- Potato starch in gelato — a lower-temperature, higher-viscosity alternative
- Tapioca starch in gelato — the cleanest-tasting of the common starches
- Butterfat vs vegetable fat — choosing which fat carries the aroma
- Maltodextrin vs glucose syrup — adding solids without adding PAC
