In middle school, I decided to spend my life’s net worth in saved-up coins. The plan pretty much carried itself out when a thallu vandi labelled ‘Arun’ in bright red letters began showing up at my school gate. For many months I bought ice cream after school almost every day, often enough to even have an open tab with the Arun anna. It always went the same way. I hand him a combination of coins totaling ₹10, a vanilla cup peels open and, by the third spoonful, the surface catches a sheen as the sun takes over. By the sixth, a rivulet escapes the rim, and then there is no going back.

There is something really inconvenient about how quickly ice cream melts in the Indian summer. Not because melting is surprising. Of course it melts. The disappointment is how little time it takes to give up. The product sold today feels like it was never designed for the place it lives in.

Ice cream here moves through heat, only briefly interrupted by refrigeration. Consider the journey of an ice cream tub. It leaves the factory in a refrigerated truck to reach the distributor, from where it goes to a petti kada¹ freezer or a street vendor’s cart, eventually to reach the customer’s home on a bike ride. Each step offers an opportunity for heat to creep in. Repeated cycles degrade its quality even if the product remains frozen. 

Counter to its name, ice cream isn’t just frozen cream. It’s a multi-phase material held together just long enough. The texture depends on the size and arrangement of these phases. 

It is impressive that we figured out this physical structure long before we had microscopes to see ice crystals or crystallography to study their growth. Remove any one of these constraints and you do not get ice cream.

Ice cream melts when one of these factors gives. Usually, it is a combination and not just one mode of failure. Ice crystals absorb heat and liquefy, fat crystals lose structure, air cells collapse, and the foam destabilizes. 

As a lover of ice cream but a hater of its melted form, I would try freezing the melted portion of my bowl, hoping to reconstruct what it once was. This never worked and always resulted in either a hard block of ice or an even weirder soup, far from what I would describe as edible, let alone appetizing.

The sharbat tradition³ grew from this and traveled the Silk Road eastward and southward for centuries. By the sixteenth century, it had reached the Mughal court, where the court dispatched riders to carry ice back to the capital for fruit sorbets⁴. The sharbat reached Europe in part through Muslim Sicily, where sugar culture took root and early frozen drink traditions first arrived on the continent. 

The classical Indian kulfi originated in 16th-century Delhi during the Mughal era, and it is made by simmering milk until it's thick and caramelized, and frozen without any churning at all. Without any overrun or foam, it has almost nothing in common (structurally) with a scoop of Häagen-Dazs and stays stable longer.

Some newer Indian brands are closely competing with the dondurma in the eccentric ice cream game. A brand called Hocco Ice Cream dropped a barfi-flavored ice cream with the snack industry mogul Haldiram’s and also a modak edition timed with Ganesh Chaturthi. Desserts like barfi ice cream come with a denser composition, making it more stable. The company also extends this thinking to products like Bijlee, which are built to handle temperature fluctuations rather than resist them. (I promise this is not sponsored. As an ice cream veteran since middle school, my expert opinion is that their product strategy is honestly intriguing. And clearly working⁷.

India’s ice cream market generates roughly ₹30,000 crore (~$3.5 billion) annually, with forty percent of that concentrated in the four months from March to June⁸. Per capita consumption has quadrupled over the past decade or so, from 400ml to 1.6 liters annually⁹. The market has grown fourfold in the same period, with projections pointing to ₹45,000 crore (~$5.25 billion) within three years, which exceeds the entire GDP of countries like Fiji or Eswatini!¹⁰

Calling something ice cream is a legal claim. To use the label, a product must be made primarily from dairy and meet minimum requirements for milk fat and milk solids. If the fat comes from vegetable oils instead, it cannot legally be called ice cream and must be sold as a frozen dessert¹¹. The distinction sounds pedantic but reflects an important formulation difference. India's ice cream map splits along a dairy line. The North and West have more concentrated dairy production, so milk fat dominates there and the ice creams are richer. Move South or East, and vegetable fats replace milk fat. They're cheaper and in heat they're more stable. Companies like Kwality Walls have historically sold products with different formulations across regions even when the branding looks identical.

No good rabbit hole ends without some level of sleuthing. To understand ice cream’s cold-chain better, of course I needed to track down as many players as I could in the Indian industry. After enough cold calls, leaders of some of the most active ice cream companies and suppliers in the market agreed to share their stories. The following section is a synthesis of conversations I had with them.

India's cold chain is built on small logistics operators with, until recently, almost no real-time visibility into temperature across the journey from factory to shelf. By the account of Varun Sheth of NOTO, India's cold chain ecosystem was fragmented when they launched about seven years ago¹². He mentioned that distributors now run their own IoT dashboards. Some brands place independent data loggers inside their shipments. Varun also pointed out that a chest freezer opened upward, lid lifted toward the face, loses far less cold air than one with a door that swings open horizontally, because cold air sinks rather than spills out. Many retail freezers in India are front-opening.

Warehouse design matters as much as temperature targets. If ice cream goes directly from -20°C storage to ambient conditions, the temperature abuse has already begun and a refrigerated truck cannot undo that.

Roli Shrivatsava, Hocco’s VP described this reality as needing to solve for India's constraints and not bypass them. They seemed to think of packaging as a part of the products and created more robust blister packaging for cones and strengthened laminated seals for their Cassata¹³.

NOTO ice cream is built around high-protein, low-sugar formulations. Their earliest formulations had extremely low fat levels due to their position as a healthier option. But in practice, it made the product significantly more sensitive to temperature shocks across the cold chain and created texture problems that consumers noticed. The fix was to increase protein levels and prebiotic fibers. Protein improves water binding and structural stability, partially compensating for reduced fat. Fiber adds body and mouthfeel that helps close the textural gap. Combined with a carefully designed stabilizer system, the product gets close to traditional ice cream texture despite lower sugar and fat. 

An elegant solution to ice cream's cold chain problem might be to stop making ice cream altogether. At least in the traditional sense. And instead, to postpone freezing until the very last possible moment.

Soft serve—or softies as they're colloquially known across India—are an entirely distinct product category from ice creams. Manufacturers heat-treat the mix rather than freeze it, which gives it a shelf life of up to nine months at ambient temperatures.

Most people eating a softie don’t realize the difference between it and a scoop out of a freezer. And why would they? The end product looks like ice cream, tastes like ice cream, and is priced like ice cream. 

The origin story of soft serve is appropriately accidental. In 1930s America, a broken-down ice cream truck sold its inventory to curious passersby who, unexpectedly, preferred the softer texture¹⁴.

Amrit Food is one of India's leading suppliers of ice cream mixes, soft serve mixes, and other specialized dairy-based beverage and dessert products to fast food restaurants.  They have supplied soft serve mixes to McDonald's since the mid-1990s. Their customer list now includes  Haldiram, IKEA, and DMart. Now they've taken this further by moving their liquid mixes into TetraPak packaging, which eliminates the cold chain requirement entirely on the upstream end. Arnav Bajaj, the Director of Amrit Food, says these commercial machines cost roughly ₹1.5–2.5 lakh ($1800–$3000) at the low end, rising to ₹10–12.5 lakh ($12,000–$15,000) for higher-end units. The upfront cost helps explain why the format remains concentrated in QSRs despite its logistical appeal.

The company’s bet is that heat treatment plus aseptic packaging will extend the shelf life even more¹⁵.The only point of transformation is the soft-serve machine at the outlet, which freezes the product at the moment of dispensing. The product doesn't exist as ice cream until the customer orders it. 

The economic drag from melty ice cream may be less dramatic than it sounds. One founder I spoke with estimated their heat-related inventory loss at two to three percent on the high side and shared that most of these partially thawed get sold anyway, rather than discarded.

The supply chain problem is going to improve over time, even as temperatures rise due to climate change, eventually paving the way for better cold chain logistics. What is not on a trajectory of getting better is the experience of eating an ice cream bar on a 42°C afternoon. That is a different kind of failure that happens in the hand. One that hardly holds it together long enough to eat.

Heat-stable ice cream is not just a future possibility. Several versions already exist, ranging from reformulated mass-market ice cream engineered in different countries to small-batch premium products in labs.

There is a widespread urban legend that ice cream ever made it to space, and 'astronaut ice cream' has since become a favorite of any space-themed gift shop. Freeze-dried ice cream is made by taking ice cream that is frozen solid and placing it in the freeze-drying machine until more than 98 percent of its water is removed by sublimation¹⁶.You end up with a brittle, chalky snack with a shelf life of three years and no cold chain requirement. It doesn't melt because there's no water left to melt. NASA solved the melting problem of ice by creating something that isn't ice cream. The resulting crumbly texture was impractical for spaceflight, and it was never on any space mission to begin with¹⁷.

A few levers can preserve the identity of ice cream while bringing it closer to a slow-melting or non-melting form. The most well-established lever is total solids content. Ice cream with low total solids simply doesn't have enough structure-building components to maintain its shape as it warms. This is how kulfi outlasts a standard ice cream bar. 

The fat network is another variable. The more interconnected the fat network, the longer the bar holds its shape. If this network is robust enough, it can hold the product's shape even as the ice phase melts away. This is why well-emulsified ice cream retains its form briefly at room temperature before collapsing.

A well-thought stabilizer system can serve as another lever. Hydrocolloids, like carrageenan, guar gum, and locust bean gum, bind water and form gels¹⁸. This increases viscosity and slows the diffusion of water to ice crystals, keeping them from growing over time. 

Strawberries and Polyphenols

An answer to non-melting ice cream already exists, and was commercially sold in Japan in six flavors. It arrived by accident when scientists at the Biotherapy Development Research Center in Kanazawa, Japan, had asked a local pastry chef to create a dessert using polyphenols extracted from earthquake-misshapen strawberries that couldn't be sold as fresh fruit¹⁹. The chef reported that dairy cream solidified almost instantly when the polyphenol was added. They realized that strawberry polyphenols work as an emulsifier, preventing the oil and water phases of ice cream from separating, helping the frozen product to retain its shape²⁰.

According to the maker, the ice cream held its form for about three hours at 40°C (104°F), while milder commercial batches, made with less polyphenol for taste, kept their shape for around an hour at room temperature. Independent tests by journalists confirmed it resists melting though it still softens over time²¹.

Bacterial Raincoats

A decade ago, a separate line of research from the Universities of Edinburgh and Dundee identified another natural protein with heat-resistant potential²².The protein, BslA, is produced by Bacillus subtilis and works by adhering to fat droplets and air bubbles, making them more stable in the mixture. A structural trick makes this unusual. BslA exists in two forms: it is soluble in water under normal conditions, but at an air-water or oil-water interface, it flips open a hydrophobic cap²³ that drives the formation of an elastic²⁴ film. This transformation stabilizes the fat-air interface in ice cream. 

The researchers likened BslA to a ‘bacterial raincoat’. Its film-forming ability stabilizes the interfaces in ice cream, helping it melt more slowly. The protein already has a foothold in the food chain. Bacillus subtilis is used to make fermented staples like natto and meju, giving it a head start on food safety regulations.

Antifreeze Proteins

This class of intervention involves antifreeze proteins, sometimes called ice structuring proteins. These are proteins produced by organisms that have evolved to survive sub-zero environments. Think cold water fish, winter wheat, certain insects, and bacteria. They adsorb directly onto the surfaces of growing ice crystals, halting further growth, while also preventing recrystallization. 

Research on ice structuring proteins found they reduced recrystallization rates by 40 and 46 percent at respective concentrations of 0.0025 percent and 0.0037 percent of total protein from cold-acclimated winter wheat grass extract²⁵. The concentrations required are extremely small.

Fish and plant-derived antifreeze proteins scale poorly for food manufacturing. A better direction is to use microbial production in the same way insulin is now produced rather than extracted from animal pancreata. Unilever commercialized this by expressing a type III ISP gene from ocean pout in genetically modified baker's yeast, and the purified protein has been used in low-fat ice cream lines²⁶.

Unilever’s Ice Cream Levers

Unilever has the most commercially serious work happening right now. The company says retail ice cream freezers account for ten percent of its value-chain greenhouse gas footprint and wants to reduce it. Their approach is to run the freezers warmer. The catch is that ordinary ice cream does not stay stable when this happens, so the product has to be reformulated to hold its structure at the warmer temperature.

Unilever tested whether its ice cream could stay stable at -12°C instead of the industry standard -18°C (0°F). Research at its Colworth R&D centre, with the Germany pilots, found that a warmer cabinet cut energy use by around 25 percent.

In November 2023, Unilever granted a free, non-exclusive license to twelve of its reformulation patents. I found no Indian startup or manufacturer that has publicly committed to heat-stable reformulation as a product strategy. The first Indian company to do it stands to define the category.

The category is already moving upmarket. IBEF²⁷describes an Indian ice cream market that now runs from affordable impulse buys to premium artisanal tubs²⁸.Hindustan Unilever (HUL)’s own presentation shows how clearly the category has already split into mass, popular, premium, and super-premium price bands. An ice cream that holds its shape for a few extra minutes but costs a bit more can easily be classified as a premium product.