Most conversations about dairy and climate treat farmers’ incomes and methane emissions as two separate problems, handled by two separate sets of institutions. One is the business of cooperatives, credit and extension. The other belongs to climate negotiators and carbon markets. On a smallholder farm, the two are the same problem, and both come down to what goes into the feed trough every morning.

This note sets out the hypothesis behind Shunya’s work, which we call the Green Horizon. It then tests that hypothesis against peer-reviewed research. Where the evidence is strong, we say so. Where it is thin, or where our own early framing ran ahead of the data, we say that too. A hypothesis that cannot survive contact with the literature is not worth building a company on.

Consider Rajeshwari Devi, 52, from a village in eastern Uttar Pradesh. She is a composite, built from the households that field surveys across the Indo-Gangetic plain describe again and again, but nothing in her day is unusual. She keeps three crossbred cows and a few goats. She is up before light to clean the shed, milk, and start the search for something green to feed. Across India, women carry most of this work; the National Dairy Development Board puts their share of dairy labour at around 70 percent, a figure we have examined in detail before.

With common grazing land shrinking every year, her cows live mostly on wheat and paddy straw, topped up with whatever concentrate the household can afford that month. The animals are underfed in exactly the nutrients that make milk: digestible energy and protein. Yields stay low, feed bills stay high, and the margin left at the end of the month is thin.

Her younger son has a college degree and little reason to stay. From where he stands, dairy looks like long hours for small money, and the city looks like the only way forward. That calculation, repeated in millions of households, is how rural economies hollow out.

Now change one variable. Every morning, a fixed quantity of fresh hydroponic maize fodder arrives at her door from a production unit a few kilometres away, grown to a consistent specification. It replaces part of the straw and part of the concentrate. Over the following months, the cows eat better, milk rises, and the feed bill per litre falls. The production unit that grows the fodder is itself a local enterprise, and it needs people like her son to run it.

That is the story. The rest of this note asks whether the science and the arithmetic support it.

Smallholder dairy is one of the largest livelihood systems on earth. The Food and Agriculture Organization (FAO) estimates that around 150 million small-scale dairy households, some 750 million people, are engaged in milk production, most of them in developing countries (FAO, 2010). India alone produced 239.3 million tonnes of milk in 2023-24 (DAHD, Basic Animal Husbandry Statistics 2024).

The same system carries a large climate footprint. FAO’s most recent global assessment puts livestock emissions at about 6.2 billion tonnes of CO2-equivalent, roughly 12 percent of all human-caused greenhouse gas emissions, with cattle responsible for 62 percent of that total (FAO, 2023). Enteric methane, produced as feed ferments in the rumen, is the single biggest part of the cattle footprint.

The regional spread in emission intensity is the key to this whole argument. FAO’s life-cycle assessment of the dairy sector estimated farm-gate emissions at about 7.5 kg CO2-equivalent per kg of fat- and protein-corrected milk in sub-Saharan Africa and 4.6 kg in South Asia, against 1.3 to 1.5 kg in North America and Western Europe, and a world average of 2.4 kg (Gerber et al., FAO, 2010). The cows of the Global South are not dirtier animals. They are underfed animals, and a large share of what they eat goes to keeping them alive rather than making milk.

Feed and fodder account for 60 to 70 percent of the cost of producing milk in India (USDA FAS, Dairy and Products Annual, 2025), and NDDB has put the figure at over 70 percent (NDDB). No other input moves a dairy household’s margin as much.

The feed that is available is also the wrong feed. ICAR’s comprehensive assessment of feed and fodder resources, reported in August 2026, estimates that under practical feeding conditions India runs a 37.5 percent deficit in concentrates and a 14.1 percent deficit in green fodder, but only a 2.4 percent deficit in dry fodder (Rural Voice, reporting ICAR, 2026). In other words, the country has roughly enough straw. What it lacks is protein and digestible energy, the two nutrients that turn feed into milk.

Straw cannot close that gap. Cereal straws average about 4 percent crude protein and 74 percent neutral detergent fibre on a dry matter basis, with organic matter digestibility of only 42 to 54 percent (Feedipedia, INRAE-CIRAD-AFZ-FAO). A cow living largely on straw spends much of her intake on maintenance. That shows up in the yield data: the average exotic or crossbred cow in India produced 8.4 kg of milk a day in 2024 and the average indigenous cow 3.54 kg (USDA FAS, 2025, citing DAHD).

Low yield hurts both ledgers at once. A cow that gives little milk earns her owner little, and she also spreads her fixed daily methane over very few litres. Fixing the diet is the one move that helps both.

Hydroponic fodder is cereal or legume seed sprouted for seven to ten days on trays, without soil, and fed whole: roots, seed residue and green shoot. The best current summary of the evidence is a 2025 systematic review of 28 controlled studies in ruminants (Vastolo & Cutrignelli, Animals, 2025). It reports average crude protein of 14.8 percent of dry matter for hydroponic barley and 12.1 percent for hydroponic maize, with far lower fibre than straw. The review found neutral to positive effects on milk yield when hydroponic forage made up 10 to 20 percent of diet dry matter, milk fat gains of 0.2 to 0.5 percentage points, and falling intake and performance once inclusion reached 40 to 50 percent.

Two features of that trial matter for smallholders. The cows were low-yielding, much like the national average, and the gains came from better digestibility of the whole diet rather than from the fodder alone. That is consistent with the 10 to 15 percent yield improvement we cite on our own platform. It is also a correction to the first draft of our Green Horizon narrative, which spoke of 20 to 25 percent gains. The peer-reviewed range is lower, and we would rather promise what the evidence supports.

Evidence from outside India points the same way and marks the limits. At a commercial dairy in Ethiopia, replacing 20 percent of concentrate with hydroponic oat left milk yield unchanged at 9.7 to 9.8 kg a day, raised milk protein and solids-not-fat, and kept net returns level. Replacing 40 percent cut yield to 8.2 kg (Fentahun et al., Dairy Science and Management, 2025). Hydroponic fodder works as part of a balanced ration. It does not replace one.

Sources: Feedipedia (straws); Vastolo & Cutrignelli (2025); ICAR assessment via Rural Voice (2026). Concentrate protein is indicative of commercial compound cattle feed.

The case for lower methane intensity rests mainly on one well-established relationship. A cow emits a baseline amount of methane simply to stay alive. When better feed lets her produce more milk, that baseline is shared across more litres, and the emission per litre falls. Gerber and colleagues showed that emissions per kg of milk decline steeply as yield rises, and that the effect is strongest where cows produce less than about 2,000 kg a year (Gerber et al., Livestock Science, 2011). Most smallholder cows in South Asia and sub-Saharan Africa fall in or near that band. FAO’s 2023 assessment reaches the same conclusion at global scale: improving productivity and efficiency is the most promising route to lower livestock emissions.

The arithmetic is simple. If a cow’s daily methane stayed flat while her milk rose by 12 percent, methane per litre would fall by about 11 percent. In practice, better-fed cows eat somewhat more, so the real reduction will be smaller than that ceiling, and it has to be measured rather than assumed.

There is also a possible second effect, from the feed itself. Diets higher in fibre tend to produce more methane per unit of feed fermented, and digestible, low-fibre feeds tend to produce less. Here the evidence for hydroponic fodder is early and mixed:

• An in vitro study at ETH Zurich found that hydroponic alfalfa and soybean seedlings cut methane by 6.6 to 6.7 percent when replacing 20 percent of a diet, and alfalfa cut it by 17.7 percent when replacing silage. Hydroponic rye increased methane, by up to 35.8 percent (Li et al., Journal of Dairy Science, 2024).
• A 2026 in vitro study with rumen fluid from Brahman bulls found that hydroponic maize harvested early (8 to 10 days) and grown with nutrient solution produced the least methane per gram of dry matter (Barros-Rodríguez et al., Fermentation, 2026).
• The 2025 systematic review found only one study that measured methane in live animals, in lambs, where emissions fell. That is not enough to support a general claim.

Water is where hydroponic fodder is strongest. Producing one kg of fresh hydroponic maize fodder takes about 1.5 to 3 litres of water, roughly 3 to 5 percent of what the same amount of forage needs under field conditions (Naik, Swain & Singh, Indian Journal of Animal Nutrition, 2015). The 2025 systematic review reports water savings of up to 90 percent. A production unit occupies a fraction of the land needed to grow the same fresh fodder in the field, and it produces through droughts, floods and heatwaves that wipe out rain-fed fodder crops.

The trade-off is also well documented. Sprouting consumes part of the seed’s dry matter; published losses range from 7 to 47 percent over six to seven days. Measured per kg of dry matter, hydroponic fodder can cost two to five times as much as the original grain (Bakshi, Wadhwa & Makkar, Feedipedia, 2017). Anyone selling hydroponic fodder as cheap feed is not being straight with farmers.

The same critical assessment is equally clear about where the technology makes sense: in developing countries with acute shortages of fodder and water, weak irrigation, high transport and fuel costs, and extreme seasonal swings in fodder prices. It notes that low-cost systems in Maharashtra brought fodder costs down to ₹2 to 3.50 per kg and cut feed cost per unit of milk by 25 to 30 percent. That describes much of rural India, and much of the Sahel, the Horn of Africa and the drylands of South Asia. The value lies in reliability, digestibility and water security, and in how well the fodder is used in the ration. We have set out the unit economics in detail in our analysis of hydroponic fodder unit cost and our comparative analysis of livestock feed.

“Doubling incomes” is an easy phrase to write, so we modelled it. The table below uses Rajeshwari’s composite household: three crossbred cows, two in milk on average, each giving 8.4 kg a day, the national average for crossbred cows. We assume a farm-gate price of ₹35 a litre and a cash cost of ₹24 a litre, of which feed is 65 percent. Family labour is not costed. These are illustrative assumptions, not field results.

The result is instructive. Better fodder alone, through yield and feed efficiency, lifts net dairy income by roughly 40 to 50 percent in this model. That is a large change for a household living on thin margins, but it is not a doubling. Getting to twice the baseline needs fertility gains, which take a full calving cycle to show, and a second income stream.

This is why we treat fodder as the foundation of a wider system rather than the whole answer. Balanced rations through Feed Right, herd records and breeding alerts through Herd Intel, and diversified livestock such as goats are all part of the same income plan.

Who gains matters as much as how much. FAO’s landmark report on women in agriculture estimated that giving women the same access to productive resources as men could raise yields on their farms by 20 to 30 percent (FAO, State of Food and Agriculture 2010-11). Reliable feed is one of those resources. When fodder arrives at the door, the hours a woman once spent cutting and carrying grass come back to her, and the animals she manages earn more.

Goats strengthen this effect. In a study of goat keeping in Meru, Kenya, researchers found that women derived and managed more income from goats than men did (Waithanji et al., Development in Practice, 2015). Goats breed faster than cattle, need less capital, eat well on the same hydroponic fodder, and can be sold quickly when a household needs cash. They cushion families against shocks such as drought, disease or a failed calving, which are common in large-ruminant dairying. For goat keepers, the economics turn on keeping kids alive, a subject we cover in why kid mortality is a profitability problem.

If hydroponic fodder only makes sense under particular conditions, the practical question is how to put those conditions within reach of a household with three cows. A tray system on every farm rarely works. It needs daily attention, consistent seed quality, hygiene against mould, and cash up front, and a single missed week wipes out the benefit.

A service model changes the economics. Fodder is grown in local production units to a common protocol, and delivered daily on subscription through the Fresh Grid network. The details of how Fodder-as-a-Service works are set out separately. Fixed costs are spread across hundreds of animals, quality is controlled centrally through Production OS, and the farmer pays for fodder she receives rather than for equipment she has to manage.

Each unit is also a rural business. Our Production Partner programme lets local entrepreneurs run fodder units, and Saarthi Partners handle last-mile delivery and farmer support. This is where Rajeshwari’s son fits in: the fodder that raises his mother’s income also gives him a skilled job without leaving the village.

The model produces data as well. Daily delivery records, milk yields and herd events captured in Herd Intel create the measurement base needed to verify productivity gains and, over time, emission reductions. That is a precondition for any carbon finance that smallholder dairy may be able to draw on in future.

The conditions that make a service model work in India are common across the Global South. Sub-Saharan Africa has the highest milk emission intensity in the world, at around 7.5 kg CO2-equivalent per kg, and some of the lowest yields, so the productivity dividend is largest there. The Ethiopian trial shows that hydroponic fodder can replace a meaningful share of concentrate in East African conditions without losing milk. In the arid Middle East and North Africa, where irrigation water is the binding constraint, the 90 percent water saving matters more than anything else.

The model has to be adapted everywhere it goes. Seed species, harvest age, inclusion rates and delivery logistics must be set for local conditions, and the rye result from Zurich is a reminder that the wrong crop choice can push methane the wrong way. We see this as an opportunity to learn from partners in every new region, and we would rather build the system with local researchers and institutions than export it unchanged. Our work with institutions is set up for exactly that kind of collaboration.

A hypothesis earns its place by naming what would prove it wrong. These are the questions we intend to answer with field data and with research partners:

• Methane in live animals. Direct measurement of enteric methane in lactating crossbred and indigenous cows fed hydroponic maize at 10 to 30 percent of dry matter, compared with matched controls.
• Yield response over a full lactation. Most trials run for weeks. We need whole-lactation data across breeds, seasons and baseline diets.
• Fertility and calving intervals. The third lever in our income model is the least documented and needs at least one full reproductive cycle to test.
• Household income, not just milk yield. Net income per household, and who in the household controls it, measured before and after service adoption.
• Women’s time. Hours freed from fodder collection, and what those hours are used for.
• Cost per kg of dry matter delivered. Published openly, so farmers and funders can compare fodder options on the basis that matters.

The Green Horizon hypothesis holds up, in a more precise form than our first draft of it. Fresh green fodder, delivered reliably and used at the right share of the ration, improves digestibility and milk yield by around 10 to 15 percent, cuts water use by an order of magnitude, and lowers methane per litre mainly by making every cow more productive. It raises household dairy income substantially. Doubling that income depends on combining fodder with better nutrition planning, herd management and diversification, and on making sure the income reaches the women who do the work.

Put together, those are not small gains. Across 150 million smallholder dairy households, they add up to a rural economy that grows, keeps its young people, and produces each litre of milk with a smaller footprint. We think that is worth building, and worth measuring honestly as we go.