Ube Gelato Recipe: Purple Yam Color, Flavor & Balance

Ube gelato is a starch problem wearing a color problem's clothes. Purple yam brings almost no sugar and almost no fat, but roughly 30% solids of nearly pure starch, so the balance has to buy back every gram of freezing-point depression the tuber fails to supply.

Ube Is Not Taro, and Not Okinawan Sweet Potato

Quick reference. Ube gelato runs 18% steamed purple yam purée, sugars 20 to 22% (sucrose 13.5%, dextrose 3.5%, lactose 4.8%), fat 6 to 7%, MSNF 9 to 10%, total solids 37 to 41%. PAC target 24 to 27, POD 16 to 19. Draw at minus 7 to minus 9 °C, 30 to 35% overrun.

Figure 1. Balance targets vs. this recipe.

Three purple tubers get sold as the same thing, and they behave nothing alike in a mix. Ube is Dioscorea alata, the greater or water yam, a true yam and the one Filipino confectionery means. Taro is Colocasia esculenta, a corm with a faint lilac fleck rather than a violet flesh, milder and more mucilaginous. Okinawan sweet potato (Ipomoea batatas, beni-imo) is a sweet potato, and it carries real sugar plus amylase activity that thins a cooked base.

The practical consequence is sugar. Sweet potato brings its own sweetness into the mix; ube brings essentially none. USDA FoodData Central lists raw yam (Dioscorea spp.) at 27.9 g carbohydrate, 1.5 g protein and 0.17 g fat per 100 g, with under 1 g of free sugars. Read that as a solids donor with a PAC of roughly zero. Every point of freezing-point depression in this recipe has to come from the added sugars and the lactose in the dairy, which is why the sugar blend below leans on dextrose harder than a fior di latte would.

Ube is also never used raw. The tuber is steamed or baked until fully tender, then peeled and passed. Beyond palatability, cooking is what gelatinises the starch, and gelatinised starch is the whole reason this gelato has the dense, faintly chewy body people recognise.

Preparing the Ube: Steam, Sieve, Weigh

Buy whole tubers where you can, frozen grated ube where you cannot. Frozen grated is the honest default outside Asia and it is already blanched, which shortens the cook. Avoid starting from ube halaya, the Filipino jam, unless you intend to recalculate: halaya is cooked down with sugar, coconut or condensed milk and butter, so it arrives carrying sugar and fat that must be subtracted from the formula rather than added on top of it.

Steam whole, unpeeled tubers for 35 to 50 minutes depending on thickness, until a skewer meets no resistance. Boiling works but leaches both color and soluble solids into the water, and you will taste the difference. Peel while warm, then pass the flesh through a fine drum sieve or blend smooth with a small portion of the recipe milk. Lumps of ube survive the mantecatore intact and read as grit in the finished scoop.

Weigh the purée cold and weigh it after cooking, not before. Steaming drives off some water and the yield varies by lot, so a formula written against raw weight drifts by several points of total solids between batches. In a shop running ube weekly, one batch of purée cooked, sieved, weighed and frozen flat in 180 g portions removes that variable entirely.

Recipe Card: Ube Gelato 1000 g

Balance targets, worth verifying against your own lot with the PAC calculator and the POD calculator:

Note where the solids come from. The ube alone contributes roughly 52 g of solids per kilo of mix, about 5 points of the 37.8% total, and almost all of it is starch that does nothing for sweetness or softness. That is the trade this flavor asks you to make.

Production Process, Step by Step

Blend the sieved ube purée with about a third of the milk until completely smooth, then combine with the remaining milk, cream, sucrose, dextrose and skim milk powder. Pre-mix the stabilizer dry with roughly five times its weight in sucrose taken from the formula so it disperses instead of clumping.

Heat to 85 °C with constant stirring and hold 30 seconds, then crash-cool to under 4 °C. The 85 °C hold clears the U.S. Pasteurized Milk Ordinance HTST equivalence of 71.7 °C for 15 seconds by a wide margin, and unlike a plain dairy base it is doing a second job: it carries the yam starch fully past gelatinisation, and it hydrates the locust bean gum, which needs the liquid phase above 75 °C to dissolve properly.

Expect the mix to thicken visibly on the way down. A starch-loaded base at 4 °C is noticeably heavier than a crema base at the same solids, and that viscosity is not a fault. Rest in maturazione at 4 °C for 6 to 12 hours, then draw in the mantecatore at minus 7 to minus 9 °C targeting 30 to 35% overrun. Move trays straight to a blast chiller at minus 25 °C for 60 to 90 minutes, then to minus 18 °C storage.

Do you still need stabilizer if the starch thickens the mix?

Yes, and less of it. Gelatinised starch binds free water and raises viscosity, which is genuinely useful, but it does not do what a galactomannan does: hold structure through the small temperature cycles a showcase inflicts on a tray every service. Starch also has a habit gums do not, covered below. The 0.4% blend here is roughly what a starch-free crema of the same solids would use, deliberately not more, because stacking starch on a full stabilizer dose is the fastest route to a gummy body.

Balancing Math: Starch Solids Without PAC

The arithmetic is short. Sucrose at 13.5% contributes 13.5 PAC. Dextrose at 3.5% contributes 3.5 × 1.9, or 6.7. Lactose from the milk, cream and skim milk powder totals about 4.8% and behaves close to sucrose, adding 4.8. That is roughly 25 PAC, sitting deliberately at the low end of the dairy window, because the starch is already contributing body and a softer mix on top of it would scoop slack.

POD lands near 17: sucrose 13.5, dextrose 3.5 × 0.74 giving 2.6, lactose 4.8 × 0.16 giving 0.8. That is a shade drier than a vanilla crema, and it is correct here. Ube's flavor is malty and faintly nutty rather than sweet, and pushing POD above 20 buries it under sugar. If your market expects the sweeter Filipino-dessert register, raise sucrose by 1 to 1.5% and accept PAC drifting to 26 or 27, which will soften the scoop slightly.

Watch the MSNF ceiling. At 9.2% there is headroom, but the temptation with a starchy mix is to add more skim milk powder for body. Above about 11% MSNF the lactose can crystallise in storage and give you sandy texture, which on a violet gelato is unusually visible.

Managing the color

Ube's purple comes from anthocyanins, principally acylated cyanidin and peonidin glycosides. Anthocyanin color is strongly pH dependent: vivid red-violet in acid, drifting to muted blue-grey as pH rises toward neutral. A dairy base sits near pH 6.6, which is exactly the range where these pigments read dull. This is why so much commercial ube product looks like a highlighter and so much honest homemade ube gelato looks lavender-grey.

Three responses, in descending order of integrity. Accept the muted tone and sell it as the real thing. Add 0.05 to 0.1% ube extract, which is standard practice in the Philippines and is where the neon in the packaged product comes from. Or nudge pH down slightly, which lifts the red-violet but fights the caseins and risks the sharp, faintly separated body low-pH dairy mixes develop. The same pigment behaviour governs anthocyanin stability in sorbet, where the natural acidity of the fruit is doing you a favor that a milk base cannot.

Variations, Service, and Storage

Ube keso folds 60 g of mild grated cheese into the cold base for the salty-sweet Filipino pairing, a combination that also lifts the perceived sweetness without more sugar. Ube macapuno adds 80 g of drained macapuno strings at the end of churning, removing 80 g of milk to hold solids. Ube coconut swaps 120 g of the milk for coconut milk, which pushes fat toward 8% and reads as a plausible base for a dairy-free version.

Serve at minus 12 to minus 14 °C, in line with other crema gelati. Storage is where this flavor differs, and it is worth telling your staff: starch retrogrades. Over several days the gelatinised amylose re-associates and the body goes from dense to short and slightly crumbly, well before any ice crystal problem appears. Rotate ube trays every 4 to 5 days rather than the 7 to 10 you might allow a nut paste flavor. The same behaviour shows up in cornstarch and potato starch bases; tapioca starch retrogrades more slowly and is the usual fix when a starch flavor must hold a full week.

On cost, ube is the dominant line by a wide margin, several times the cost per kilo of any domestic tuber and more still if you buy halaya rather than frozen grated. Price it as a specialty flavor rather than slotting it beside fior di latte, and let the color do the merchandising work in the showcase.

Related Concepts

• Starch Retrogradation in Gelato, why starchy flavors go short
• Tapioca Starch in Gelato, the slower-retrograding alternative
• Anthocyanin Color Stability, the pigment chemistry behind the grey
• Sucrose vs Dextrose, the PAC and POD trade
• Yogurt Gelato Recipe, another high-solids dairy base
• How to Balance a Gelato Recipe, the master method