Indian agritech spent a decade learning one expensive lesson: the farmer is the user, not the payer. The companies that survived the funding winter are the ones that stopped selling to farmers and started monetising the transaction around them.
The reset is in the numbers. Agritech funding fell to $241 million across 60 rounds in 2025, from $392 million across 105 rounds in 2024 - a 39% drop, per Tracxn. But that headline hides the real story, which is compositional: the asset-light input marketplaces and advisory apps didn't just slow down, they largely disappeared. What's scaling now is full-stack. The reason is a unit-economics wall every founder in the sector hit at the same spot.
Here's the wall, and how to read it if you're building or studying one of these:
- Revenue per farmer is structurally tiny, so it's the wrong unit entirely. India has 140 million-plus smallholder households and under 1% tech penetration. A model whose economics depend on average revenue per farmer is dead on arrival - the number can't get big enough, fast enough, to cover a CAC that stays stubbornly high because trust-building in rural markets is slow and manual. Stop modelling ARPU. Model contribution per transaction and lifetime financing spread.
- The money is in the adjacent layer, not the app. The one agritech that's fully profitable at scale doesn't make it selling software - it makes it on warehousing and an NBFC arm doing thousands of crores in commodity financing. That's the tell for the whole sector: the defensible margin sits in credit, aggregation, and supply-chain control, using the platform's transaction data as the underwriting edge. The app is customer acquisition. The balance sheet is the business.
- Working capital is the metric that kills the full-stack winners. The moment you go full-stack - holding inputs, aggregating output, financing the farmer - you take on inventory and receivables, and a 'fast-growing' agritech can quietly be a working-capital furnace. In diligence, inventory days and DSO tell you more than GMV ever will. Growth funded by an expanding working-capital gap is not growth; it's a loan you're giving your own P&L
So the question that sorts a real agritech from a subsidised one: strip out the financing income and the government/enterprise contracts - is there a profitable business left that farmers actually pay for? Usually there isn't, and that's fine, as long as the model knows it and monetises the layer that does pay
#Agritech #UnitEconomics #DueDiligence #AgriFintech #StartupIndia #NBFC #Credit #Farmers #India #Bharat #Consulting #Growth
19 August 2026 à 05h30
View on LinkedIn →
