Investing in Agriculture Begins With the Farmer

A farmer commits money months before knowing what the weather will allow or a buyer will pay. When tomatoes flood the market, the farm-gate price collapses while the price of ketchup barely moves. Processing, storage, distribution and branding explain part of that difference. They also reveal who can wait, who can preserve value and who must sell immediately.

Agricultural investment should improve that equation by starting with the farmer's needs: more dependable harvests, products that work, less uncertainty and risk, and a fairer share of the rewards. Three priorities follow — each obvious in theory, missed in practice.

1.Better crop varieties are agriculture's most underleveraged investment.

Every rupee invested in agricultural research returns ₹13.85 in wider economic benefits. That figure, cited in the Economic Survey 2023–24, is not a development aspiration — it is a documented return.^[1] Yet public agricultural research spending sat at 0.49% of agricultural GDP in 2020–21, rising to just 0.54% when private spending is included. ICAR–NIAP recommends reaching 1% by 2030.^[2] We are not close.

Private capital has made the same choice. In 2023, India's agrifoodtech sector raised $940 million. Agricultural biotechnology — the category that includes seed genetics, biological inputs and animal health — received $4 million of it, or 0.4%. E-grocery received $420 million.^[3]

In 2023, agricultural biotechnology and farm robotics together received just $7 million — less than 1% of the total.^[3] In 2024, Zepto alone accounted for roughly 56% of India's $2.5 billion agrifoodtech funding.^[5]

The gap is not accidental. Biotech takes time: Agrico (a Dutch cooperative) started with 200,000 seedlings and waited nine years for three commercial varieties.^[6] Exit paths are thin. IP is hard to protect in a fragmented seed market. Consumer-facing businesses offer faster signals and cleaner returns. Capital responds rationally to these structures. Changing where capital goes means addressing these structures — not just pointing to the gap.

The returns on better varieties show up most clearly when conditions turn bad. In Odisha, a J-PAL evaluation found that farmers who planted Swarna-Sub1 — a submergence-tolerant rice variety — earned ₹2,969 more per hectare in a flood year than other farmers in the two-year study, with no yield disadvantage in non-flood years.^[7] The variety also changed behaviour: farmers invested more in inputs and improved cultivation practices, accounting for 41% of the long-term rice yields increase.^[7] Drought tolerance works the same way. Sahbhagi Dhan, developed by IRRI and CRRI, yields 4–5 t/ha under normal conditions against around 2.5 t/ha for traditional long-duration varieties; in a severe drought year it still yields 1–2 t/ha where traditional varieties often yield nothing.^[8] Pest and disease resistance reduces costs: fewer fungicide and pesticide applications, lower input bills, and a higher share of export-quality output from the same land and labour.

These effects are not marginal. A variety that protects yield when floods come, maintains an advantage when they don't, enables an additional crop through shorter maturity, and cuts fungicide spend year-round compounds across every season.

Sahyadri Farms illustrates this at the commercial level. It tests imported grape varieties through R&D farms supported by soil, water and residue laboratories.^[9][10] During Maharashtra's unseasonal rains in 2025, the wider grape industry suffered 60–70% crop losses while Sahyadri's ARRA plantations saw limited impact.^[11] Grapa (Sahyadri's partner), which commercialises the ARRA varieties, reports 90% export-quality fruit for ARRA Sweeties versus 60% for Thompson Seedless.^[12] These examples illustrate what better varieties can mean for a farmer's income in a bad year.

Agrico's (a Dutch cooperative) potato breeding programme illustrates the same logic. Its Next Generation varieties resist late blight, supporting more dependable yields and organic cultivation.^[13] Another variety Levante also resists golden potato cyst nematode.^[14] The breeding programme ends with grower representatives choosing which varieties go forward, closing the loop between the lab and the field.^[6]

In India, ICAR had developed 455 stress-tolerant varieties in 2024 alone — 92 targeting extreme abiotic stress including drought, floods, waterlogging and heat.^[15] Since 2014, nearly 3,000 climate-resilient varieties have been released across field and horticultural crops.^[16] Climate-resilient wheat varieties covered more than 60% of India's wheat area in 2024–25.^[17] The Economic Survey projects a 20% decline in rainfed rice yields by 2050 without adaptation.^[18] Much of that loss is preventable with varieties that already exist. The gap is not in the science — it is in carrying what has been developed through to the farmer at scale.

SV Agri illustrates what that bridge requires: screening Agrico's varieties for Indian climates and buyer requirements, multiplying seed through aeroponics and field production, then providing cultivation support and market access.^[19] There is an opportunity to back more businesses doing this — developing improved varieties, testing them under local conditions and building the seed-production and market-access infrastructure needed for adoption at scale. The goal is to create commercially viable businesses that deliver more dependable harvests and incomes.

2.Farmers are not beneficiaries. They’re customers. Design accordingly.

The NSO's 2018–19 survey put average monthly agricultural household income at ₹10,218. Net crop receipts contributed just ₹3,798 of that — 37.2%. Crops and livestock together contributed 52.7%; wages accounted for 39.8%.^[20] In that context, a failed seed or input wastes not just money but months of labour, irrigation and borrowed capital. Replacing the packet cannot replace the harvest.

The high stakes make quality a precondition, not a feature. Bio-inputs show what happens when it is absent: government laboratories found 80 of 111 biofertiliser samples tested during January–September 2022 substandard — about 72% of that tested set.^[21] Once a farmer loses confidence in a category, even effective products struggle to earn a fair hearing. Building that confidence requires transparent claims, demonstrations under local conditions and a genuine choice about what to use. Adoption must be earned, not forced. In July 2025, the Agriculture Minister directed states to stop compulsory bundling of nano-fertilisers or biostimulants with conventional fertilisers.^[22] Compelling farmers to buy additional products undermines both choice and trust.

The same logic applies to equipment. In 2023–24, agriculture, forestry and fishing employed 64.4% of working women nationally and 76.9% of rural working women.^[23] A product mindset starts by observing what those women actually do: testing weight, grip, operating effort and safety with them, not for them. In a Chhindwara trial with 20 women, a twin-wheel hoe reduced perceived drudgery by 26.6% and halved labour requirements.^[24] Designing for the actual user is not a social objective — it is what makes the machine useful.

It also means designing for the actual farm. Agriculture Census 2015–16 put 86.1% of operational holdings below two hectares.^[25] In an ICAR programme in Goa, farmers wanted reapers with cutting widths of 60 cm or less because 120-cm models were simply too unwieldy.^[26] Mahindra's Yuvraj 215 NXT NT and John Deere's narrow 3028EN show that adaptation is possible.^[27][28] The harder question is whether farmers can afford, access and maintain what is built. VST's February 2024 earnings call put power weeders at ₹40,000–70,000 and reapers above ₹1 lakh, and noted purchase delays due to farmers' cash-flow constraints.^[29] Hiring and shared ownership models can help, but only where demand is dense enough to keep utilisation viable.

Lastly, affordability means little if the product stops working and no one can fix it. The drone programme shows what happens when delivery outpaces service. A government-cited study of 500 drones under Namo Drone Didi found 43% of operators facing transport problems and 69% reporting costly vehicle hire.^[30] A March 2026 parliamentary reply cited crashes, overheating batteries and repair delays.^[31] In one instance, a user waited 10–15 days for a technician after a GPS fault. When her battery failed, the ₹40,000–50,000 replacement cost was beyond her means. As of the May 2026 report, the drone had been idle since December 2025.^[32] A technically capable machine stopped being a usable service.

A product business that ignores service is not selling a product — it is selling a promise that fails at first contact with the farm. Local technicians, readily available parts and repair commitments that fit the short windows of farm operations are not optional features. For the farmer, quality means work completed on time at a cost the household can bear.

3.Risk-sharing isn't charity. It's good business design.

When farmers don't adopt a new variety, input or practice, the standard explanation is that they are conservative or slow to change but they are just minimizing risk.

A farming household faces heat, unpredictable rainfall, pests, input quality failures, labour shortages, price collapses and payment delays — often simultaneously. Working capital is committed months before returns arrive. A new variety or machine adds another unknown: will it work here, and can the family survive a failed season? Holding back is not irrationality. It is an accurate assessment of who bears the losses if something goes wrong. Change the risk structure, and decisions change too.

Evidence from Andhra Pradesh makes that link direct. In a 2009 experiment covering 1,479 farming households, providing rainfall insurance increased the likelihood of planting higher-return cash crops by six percentage points — before any claim was paid.^[33] The cover itself changed the decision. The study found no statistically significant increase in total investment, and effects were concentrated among more educated farmers. But the direction is clear: reduce the downside, and farmers take the opportunity.

Part of that downside is informational — not knowing whether a practice will work in their conditions. Farmers actively seek guidance; in a Punjab study, 66% of 50 respondents reported always using YouTube for agricultural advice.^[34] The appetite is real. But information alone doesn't shift the risk when the farmer absorbs the full cost of a failed recommendation. Advice becomes credible when those giving it also share the consequences of being wrong. AI tools can widen access — voice support, local languages, crop-specific guidance, as Digital Green's FarmerChat illustrates across six languages^[35] — but access to advice only matters if the advice is correct, locally calibrated and connected to reliable inputs, finance and buyers.

Ultimately, the risk structure is set by how the business is designed. Sahyadri shows one approach. The company combines agronomy, testing, packhouses, processing and access to domestic and export buyers.^[9][36] For FY2023–24 it reports procuring and processing over 300,000 tonnes of horticultural produce from more than 26,000 farmers,^[37] with Kissan ketchup production for Hindustan Unilever as one downstream use of that infrastructure.^[38] The result is that farmers have somewhere to sell, at predictable prices, across multiple crops.^[10]

But the structural element that matters most is ownership: 22,500 farmer-members hold ownership with voting rights.^[10] Farmers have a stake in downstream value, not just the farm-gate price. Adopting a new variety or meeting a quality standard becomes less daunting when you own a share of what that quality creates. An RBI working paper estimates farmers receive 33% of the consumer rupee for tomatoes, 36% for onions, 37% for potatoes.^[39] NABCONS estimated post-harvest losses of 6–15% for fruits and 5–11% for vegetables.^[40] Processing, storage and farmer ownership of downstream value can improve both numbers. The question is whether the businesses being built are designed to do that.

Conclusion

Every part of India's food system — the processor, the trader, the retailer, the consumer — depends on the farmer having a reliable harvest to start with. When the tomato crop fails, ketchup factories run short. When wheat yields fall in a heat year, the deficit shows up in prices months later across the country. The farmer is not one stakeholder in the food system. The farmer is its foundation.

Yet the system is organised as if the farmer's resilience can be assumed. Investment flows to the parts of the value chain that can protect their margins — processing, branding, distribution — and away from the part that cannot: the farm itself. The result is a food system that is more fragile than it needs to be, built on a base that absorbs all the weather risk, all the input uncertainty, and most of the price volatility, with the least capacity to do so.

India has the scientific capability and entrepreneurial talent to make agriculture and the wider food system work for everyone. The starting point must be the farmer's priorities: research that protects harvests, products backed by dependable service, and business models that share risk and reward fairly. When farming offers a more dependable livelihood, farmers gain the confidence and capacity to invest — in better varieties, in better inputs, in their land. That investment strengthens the base on which the rest of the food system rests.

The knowledge is there. The examples are there. What is needed is the willingness to direct capital, product design and business structure toward the farmer's interests — not as a social obligation, but as the condition on which a functional food system depends.